Fixed Issues

This release provides fixes and enhancements to pre-existing functionality:
capm370
This release provides fixes and enhancements to pre-existing functionality:
3.7.17 Fixes
  • Symptom:
    Vulnerability tests against
    https://<
    Data Aggregator Host
    >:8582
    show the
    LUCKY13
    and
    Secure Client Renegotiation
    vulnerabilities.
    Resolution:
    The
    LUCKY13
    and
    Secure Client Renegotiation
    vulnerabilities on Data Aggregator Host when in https mode have been resolved.
    (3.7.17, DE451369)
  • Symptom:
    When a data collector upgrade that has been initiated from the user interface (UI) has failed, the data collector will remain upgrading.
    Resolution:
    Update data collector status to failed upgrade if a registration message is not received within 10 minutes of initiating an upgrade through the UI process.
    (3.7.17, DE470798)
  • Symptom:
    When a data source synchronizes devices that are considered the same based on Performance Center reconciliation, a SQL exception can be thrown about duplicate key. Also, the same query is slow due to having to convert columns between ISO-8859-1 and UTF-8, and missing an index.
    Resolution:
    Updated the query failing to use DISTINCT to handle the duplicate keys. Added new mysql index, and encoded the temporary tables using UTF-8, to speed up queries.
    (3.7.17, DE473120)
  • Symptom:
    The
    dr_validate.sh
    script requests linux user password during database connectivity testing.
    Resolution:
    Modified database connectivity check to use existing passwordless ssh.
    (3.7.17, DE477938)
  • Symptom:
    Output of some Open API queries in CSV format may include extra columns and device information may be missing in some rows.
    Resolution:
    With this fix, output of Open API queries will not have extra empty columns and device information would be correctly repeated when necessary.
    (3.7.17, DE478981)
  • Symptom:
    vEdge Interfaces that have non-unique private IP addresses are being incorrectly reconciled within the
    DX NetOps Virtual Network Assurance
    (VNA). This is leading to missing interface and tunnel items and also interfaces and tunnels being associated with the wrong devices.
    Resolution:
    The Viptela Plugin has been updated to no longer use the private IP of an interface as a means to uniquely identify it. This will allow the missing interfaces and tunnels to be created, and will resolve the issue with items being associated with the wrong devices.
    (3.7.17, DE479272)
  • Symptom:
    Unable to identify a user's permission group through REST to be able to set the user's Default Group to My Assigned Groups.
    Resolution:
    Add permission group information to user REST endpoint.
    (3.7.17, DE479547)
  • Symptom:
    Can't add items to groups within the User Defined Groups under the
    Network Flow Analysis
    (NFA) data source inventory in Performance Center.
    Resolution:
    With this fix, users can add items to groups within the User Defined Groups under the NFA data source inventory in Performance Center.
    (3.7.17, DE479574)
  • Symptom:
    On a Performance Center that has multiple IP addresses, Performance Center will select the first interface and first non-loopback IP for the Performance Center Service default Web Site Host value. This can lead to unreachable Performance Center IP used by Event Manager and users/services that contact the Performance Center service.
    Resolution:
    Update Performance Center to set the default Web Site Host by trying to determine the fully qualified hostname of Performance Center. If a fully-qualified host name (FQHN) cannot be determined, it will fallback to first interface/first non-loopback IP. These can still be overridden via SsoConfig to set Remote Value for Performance Center Web Site Host.
    (3.7.17, DE479930)
  • Symptom:
    The Cisco Application Centric Infrastructure (ACI) Console fails to load the topology when selecting certain items.
    Resolution:
    The ACI Console logic has been updated to more gracefully handle incorrectly formatted Open API responses.
    (3.7.17, DE480372)
  • Symptom:
    The Admin group UI allows users to modify locked groups by adding or removing items.
    Resolution:
    With this fix, the Admin group UI does not allow to modify locked groups by adding or removing items.
    (3.7.17, DE481382)
  • Symptom:
    When new items come in to poll, if there is a bad polled item ID (NULL) in the request, a NullPointerException would be thrown and it would fail to start polling the additional item ID(s) after the NULL item ID. It will never schedule those polled item IDs.
    Resolution:
    Updated start polling code to handle if NULL are passed it, and print a WARN message and continue processing the non-NULL item IDs.
    (3.7.17, DE482423)
3.7.16 Fixes
  • Symptom:
    Python2 is EOL, and there are a few admin/debug python 2 scripts shipped with Data Aggregator/Repository.
    Resolution:
    Updated the Python scripts to run against Python 3. Python 3 installation in Linux is now required to run these scripts.
    (3.7.16, DE440532)
  • Symptom:
    The procedure for configuring daproxy and consul services as HTTPS on a Fault Tolerant CAPM is incomplete and confusing. Configuring daproxy service from the Proxy server to the data aggregator's yields certificate_unknown errors.
    Resolution:
    The traefik service has been upgraded to version 2.2.8 to resolve certificate errors between the daproxy service and the backing data aggregator's. The documented procedures for configuring FT services as HTTPS have been rewritten and clarified.
    (3.7.16, DE460814)
  • Symptom:
    On a system configured for Fault Tolerance, /odataquery points to the active data aggregator instead of data aggregator proxy.
    Resolution:
    With this fix, on a system configured for Fault Tolerance, /odataquery now points to the data aggregator proxy.
    (3.7.16, DE471203)
  • Symptom:
    When a poll value delta is calculated to be greater than max positive 32 or 64 bit integer,
    Performance Management
    only drops the poll response if it happens during a rollover. This happens when a Simple Network Management Protocol (SNMP) agent returns an nearly impossible delta that could happen within the poll cycle.
    Resolution:
    Changed the SNMP delta calculation code to handle these very large deltas even when not in a rollover situation. The max delta is controlled via configuration. See docs if you want to increase or decrease the large 32 or 64 bit delta value used in this check.
    (3.7.16, DE472708)
  • Symptom:
    Removing Proxy port field from Viptela Config is causing NullPointerException (NPE) in
    server.log
    .
    Resolution:
    Added additional checks for NPE in VNA
    server.log
    .
    (3.7.16, DE473370)
  • Symptom:
    An insufficiently sized thread pool leads to a WARN message to be logged for a mis-fired Poll Item or Group ETL job every hour when the DIM Item ETL is running at the same time.
    Resolution:
    The ETL scheduler's thread pool size was increased and made configurable.
    (3.7.16, DE474462)
  • Symptom:
    "No data to display" after upgrade to 3.7.14.
    Resolution:
    Added ETL health check to the
    dr_validate.sh
    and
    etlHealth.sh
    scripts to detect ETL issues before upgrade.
    (3.7.16, DE474520)
  • Symptom:
    Multiple records with the same item_id and facet_id value pairs are present in the item_facet table in the Data Repository. There are internal workflows in the data aggregator that can result in the same facet being added to the same item in a single database request.
    Resolution:
    The item repository data base layer now removes duplicate item_facet records when inserting a single batch.
    (3.7.16, DE475453)
  • Symptom:
    After running the Performance Center disaster recovery script, the new Event Manager IP may not be sent to data aggregator due to timing between EM synchronization and event poller threads.
    Resolution:
    Updated Performance Center disaster recovery script to update the DB field, used for the EM URL sent to the data aggregator during re-registration, with the new EM IP address. This allows the event poller thread to see it on startup and send the correct EM URL to data aggregator.
    (3.7.16, DE475891)
  • Symptom:
     VNA orgVsConnectedDevicesFile during Versa Inventory poll is not getting updated.
    Resolution:
    Fixed the issue ,now the file will get updated during the inventory poll.
    (3.7.16, DE476346)
  • Symptom:
    Odata count query on the metricfamilyhistory entity with a filter may return incorrect count of entities.
    Resolution:
    With this fix, odata count query on the metricfamilyhistory entity with a filter will return correct count of entities.
    (3.7.16, DE476454)
  • Symptom:
    Non-application users are able to login to the Performance Center MySql database without providing a password.
    Resolution:
    Update Performance Center install script to remove anonymous user from the MySql database.
    (3.7.16, DE476790)
  • Symptom:
    OOB groups under all old tenants displayed old group names.
    Resolution:
    Added upgrade script to update OOB group names of groups still using old group names.
    (3.7.16, DE477368)
  • Symptom:
    ETL failures may occur if multiple data aggregators are or previously have been running simultaneously while using the same Data Repository schema instance. This will result in NetOps Portal dashboards showing no data or not showing data for recently discovered items. This may occur by accidentally using the same Data Repository information when installing two data aggregators (that are not intended for Disaster Recovery) or when migrating the data aggregator to a new system and both the old and new data aggregator are running and the Data Repository is not blocking access to the old data aggregator.
    Resolution:
    To protect the Data Repository, the data aggregator will now verify that no other data aggregator is attached to the same Data Repository schema instance during the data aggregator startup process. If a data aggregator detects that another data aggregator is actively using the Data Repository, it will shut down immediately and log information to the
    shutdown.log
    file as to why it has shut down. A long-running data aggregator will take precedence over a newly started data aggregator, and the newly started data aggregator will shut down while the long-running data aggregator remains running.
    (3.7.16, DE477749)
  • Symptom:
    After customizing a view any attempts re-edit the view will not launch view settings dialogue until after the page is refreshed.
    Resolution:
    Addressed issue that after editing a view the user is unable make subsequent edited to view configuration until after the page is refreshed.
    (3.7.16, DE477881)
3.7.15 Fixes
  • Symptom:
    Old RIB document files can build up in the Performance Center /tmp/ribcache.
    Resolution:
    Performance Center now cleans up the old version(s) of a RIB document file when a new version is available.
    (3.7.15, DE452506)
  • Symptom:
    REST endpoint requesting the SNMPv3 EngineID every time returns a new SNMPEngineID that is not does not much one used in traps.
    Resolution:
    With this fix, REST endpoint requesting the SNMPv3 EngineID would return correct SNMPEngineID that does much one used in traps.
    (3.7.15, DE457958)
  • Symptom:
    8443 HTTPS is enabled by default in VNA, though https is not supported.
    Resolution:
    Disabled HTTPS port on wildfly.
    (3.7.15, DE462717)
  • Symptom:
    On-Demand, view option 'Chart per item with multiple metrics' when applying Metric Calculate level by Device the sort order is not being honored.
    Resolution:
    Addressed issue with On-Demand, view option
    Chart per item with multiple metrics
    when applying Metric Calculate level by Device the sort order is not being honored.
    (3.7.15, DE464964)
  • Symptom:
    When installing/upgrading the Performance Center, data aggregator or data aggregator Proxy, and there is 1 private IP address and 1 public IP address on the machine, the installer does not prompt for which IP address to use for consul to bind and listen on. It instead just uses the private IP address.
    Resolution:
    Updated the installers to ask if there is 1 private and any number of public IP addresses.
    (3.7.15, DE467204)
  • Symptom:
    When running Alarm Console against
    DX NetOps Spectrum
    10.4.x+ (20.2.x), it may time out or throw an error when not using a custom group. This is because the landscape group synchronized from
    DX NetOps Spectrum
    to Performance Center changed and Alarm Console does not take that into account. So it is using Global Collection groups, and if no GCs are defined, all synced Spectrum items.
    DX NetOps Spectrum
    also has a limit on 150 comparisons in the webservice used for Alarm Console, so if too many GCs are passed, it will fail.
    Resolution:
    Updated the logic to check for the old and new Spectrum landscape group IDs. By using landscape group, Spectrum can return all events for the landscape group members quicker than specifying all items in groups. In addition, to resolve the issue with 150 comparisons,
    Performance Management
    only adds up to 150 comparisons to the webservice request, and prints a message to the performance center
    service.log
    file when more than 150 comparisons is hit.
    (3.7.15, DE468346)
  • Symptom:
    During installation of the data repository, the I/O scheduler is always set to the same value regardless of type of storage.
    Resolution:
    Update the data repository installation process to allow the I/O scheduler to be updated to a different value based on storage type.
    (3.7.15, DE468411)
  • Symptom:
    During Performance Center and DR install, the disaster recovery scripts that may have been customized are overwritten without a backup.
    Resolution:
    Update the Performance Center and DR installers to make backups of the disaster recovery scripts before setting down updated versions.
    (3.7.15, DE468863)
  • Symptom:
    The Open API may not return all groups in expand clause for queries starting from configuration entity types (devices,interfaces,etc.) if the total number of groups is more then a hundred.
    Resolution:
    With this fix, the Open API would return all groups in expand clause for queries starting from configuration entity types (devices,interfaces,etc.) regardless of the number of groups.
    (3.7.15, DE470008)
  • Symptom:
    With multiple columns trend chart views, the Y-axis title is partially hidden when metric unit is scaled to numeric values that forces axis height to 1000 units.
    Resolution:
    Addressed the partially visible Y-axis title when numeric values forces axis height on trend chart to 1000 units seen with multiple columns trend chart views.
    (3.7.15, DE470115)
  • Symptom:
    The System Status in Performance Center UI does not show the backend data aggregator's - only shows daproxy - on a system configured for fault Tolerance, when the consul service on DAproxy is configured for HTTPS.
    Resolution:
    Performance Center was not recognizing the consul service is configured for HTTPS when calling consul rest services. Performance Center and data aggregator have been updated to recognize when FT's consul is in HTTPS mode, so data aggregator's can be shown in System Status.
    (3.7.15, DE471182)
  • Symptom:
    Site Context Pages rendering business hours in subtitle when business hour filter is disabled.
    Resolution:
    Fixed issue seen on Site Context Pages with business hours rendered in subtitle when business hour filter is disabled.
    (3.7.15, DE471214)
  • Symptom:
    The data aggregator Proxy is unable to write to
    daproxy.log
    .
    Resolution:
    Update the owning group of the log directory for data aggregator Proxy.
    (3.7.15, DE471368)
  • Symptom:
    DAProxy installer fails due to missing jre folder.
    Resolution:
    Update consul server installer to not cleanup extra jre folder on upgrade.
    (3.7.15, DE471583)
  • Symptom:
    When device is discovered that are related to F5 metric the context tabs for F5, available on Switch Context page, are not available on the Device Context page. When preference is to not associate the device as a switch device, requires manual creating F5 Tabs on Device Context page.
    Resolution:
    When F5 related devices are discovered as generic device and prefer to not associate the device as a switch. Provide an automated script to import context tabs for F5, available on Switch Context page, to the Device Context page. Updated to better handle view suppression within the Device Context Page.
    (3.7.15, DE471952)
  • Symptom:
    When a role for a user with My Custom Groups was changed via the REST API, the My Custom Groups group was deleted.
    Resolution:
    Update the role REST API call to pass down the permission group ID for the user. A mismatch for the new permission group was causing the deletion.
    (3.7.15, DE472070)
  • Symptom:
    When creating notification emails, the subject and body fields do not behave correctly (they take forucs from other fields, do not correctly respond to clicking the Insert button, and do not maintain changes made to their content).
    Resolution:
    When creating notification emails, the subject and body fields now behave correctly.
    (3.7.15, DE472436)
  • Symptom:
    Cisco ISPLA polling can fail with NullPointerException if component IndexList has unexpected format for a given AttributeGroup.
    Resolution:
    Cisco IPSLA poller gracefully handles the unexpected component IndexList format, log an appropriate message, and ensure data from the other valid AttributeGroups can be collected.
    (3.7.15, DE472770)
  • Symptom:
    InnoDB fails to initialize during install. A common symptom is that functions that should have installed into MySql did not. Confirmation can be found in the MySql log.
    Resolution:
    Update the installer to check the fs.aio-max-nr setting and update it to a higher recommended value.
    (3.7.15, DE472860)
  • Symptom:
    A null pointer exception is recorded in the VNA logs and polling does not acquire a full inventory.java.lang.NullPointerException at com.ca.em.sdn.gateway.plugins.viptela.Devices.lookUpEntity(Devices.java:359) [viptela-plugin.jar:].
    Resolution:
    Changes were made to handle this exception w/out disrupting/aborting vManage polling.
    (3.7.15, DE472882)
  • Symptom:
    When generating CSV reports for trend chart views the CSV Export of data does not reflect Business Hour periods in both raw and scaled format.
    Resolution:
    Export of data when generating CSV reports for trend chart views data dis not reflect Business Hour periods in both raw and scaled format.
    (3.7.15, DE473237)
  • Symptom:
    ClassCastException in the
    Expression.log
    file.
    Resolution:
    Improved error handling in mvel snmpProtectedDiv function, and improved expression logging.
    (3.7.15, DE473263)
  • Symptom:
    End of VNA installation the ReadMe is still pointing to docops.ca.com.
    Resolution:
    Updated to point to the correct techdocs.broadcom.com location.
    (3.7.15, DE473504)
  • Symptom:
    While installing VNA the end timestamp is displaying as dd_mm_yyyy_hh_mm_ss instead of actual timestamp.
    Resolution:
    Display the actual end timestamp.
    (3.7.15, DE474261)
  • Symptom:
    There is no indication of what version of CE is installed.
    Resolution:
    Update the CE installer to write the version when installing.
    (3.7.15, DE474889)
  • Symptom:
    Dynamic Trend View is not properly suppressed on Device Context Page when configured within the Dashboard Editor.  .
    Resolution:
    Addressed issue that the Dynamic Trend View is not properly suppressed on Device Context Page when configured within the Dashboard Editor.
    (3.7.15, DE474985)
3.7.14 Fixes
  • Symptom:
    OData API is fetching old metric data when start time and end time equals ZERO.
    Resolution:
    OData API will return empty data when start time and end time equals ZERO.
    (3.7.14, DE442822)
  • Symptom:
    The data aggregator failed over to the standby data aggregator with no clear indicator in the logs about why the decision was made to fail the over to the standby.
    Resolution:
    Connection and read timeouts have been added to the data aggregator ActiveMQ and REST status checks. Additionally, the Consul Extension logging has been updated to be more verbose when a system that is currently the active data aggregator goes down and a fail-over occurs.
    (3.7.14, DE448360, 31918553)
  • Symptom:
    When TransientDBConnection logging is set to debug, SQL statements with passwords are logged.
    Resolution:
    Update logger to sift out SQL with passwords.
    (3.7.14, DE456484)
  • Symptom:
    The Group Editor's Group Description gets stuck on the invalid description if over 255 chars long.
    Resolution:
    The Group Editor allows the user to modify a description that is over 255 characters.
    (3.7.14, DE458149)
  • Symptom:
    When discovering Check Point firewall virtual systems with context support, many virtual system items were duplicated with empty context names.
    Resolution:
    Enhanced the virtual system validation process not to discover the virtual systems that are improperly configured with empty context names.
    (3.7.14, DE461334)
  • Symptom:
    The data aggregator is unresponsive to RIB, REST, or any other request, but java memory usage/garbage collection is fine. Could be a deadlock situation between a couple tasks the data aggregator is running. There is a deadlock possibility when doing a REST/UI delete, change detection, discovery profile running, and an OData call that filters on groups.
    Resolution:
    Updated the OData filter code to negate the possibility of the deadlock situation.
    (3.7.14, DE464968)
  • Symptom:
    odata query may not return device information in expand when it starts from metric family/component for example portmfs?$expand=device.
    Resolution:
    with the fix odata query will return device information in expand when it starts from metric family/component for example portmfs?$expand=device.
    (3.7.14, DE465743)
  • Symptom:
    Open API in some circumstances may not honor tenants boundaries.
    Resolution:
    With this fix, the Open API should honor tenant boundaries all the time.
    (3.7.14, DE466422)
  • Symptom:
    When discovering Cisco switch's Environment sensors, only one of the relevant metric families is discovered; or, all of the environmental sensors metric families, but some have no components.
    Resolution:
    Improved discovery logic for Cisco Environmental Sensor vendor certs. Added new MVEL method that allows discovery expressions to scan all table entries for an OID, instead of only the first entry.
    (3.7.14, DE466858)
  • Symptom:
    On the System Health page, when you page to the next page in a table the section is collapsed each time.
    Resolution:
    Instead of setting the collapsed property to be purely based on the status of the section, we're also taking into account the current collapsed state.
    (3.7.14, DE467043)
  • Symptom:
    When proxying as a user who uses a different Locale, the Group Rule Editor shows incorrect translations in its group selector.
    Resolution:
    Modified group selector to get locale from active user.
    (3.7.14, DE467052)
  • Symptom:
     Scheduling "All Pages of multiple page views" for On-Demand report option "Per metric by single item" the PDF output combines multiple metrics the trend charts.
    Resolution:
    Fixed issue when scheduling On-Demand report option "Per metric by single item" using "All Pages of multiple page views" report the PDF output combines multiple metrics the trend charts.
    (3.7.14, DE467175)
  • Symptom:
    The responsiveness of the ACI Console dashboard has been improved by updating the ACI Console to request data for multiple items at a time. This should prevent the dashboard from taking excessive time to load when a large number of switches are part of the topology.
    Resolution:
    The ACI Console can take up to a minute to respond after clicking on one of the icons in the hierarchy. The extended time that it takes to load the topology is ultimately based on number of switches that show up in the topology view on the right.
    (3.7.14, DE467188)
  • Symptom:
    The ACI console displays non-ACI User Domains in the hierarchy.
    Resolution:
    The ACI console has been updated to display only user domains that contain ACI technology group. By having the user domains in data aggregator store the information about the technology groups, Performance Center can directly fetch user domains that contain ACI.
    (3.7.14, DE467199)
  • Symptom:
    The initial state of the sections displayed in the new Administration Group Editor was confusing as some sections were initially uncollapsed and some sections were collapsed.
    Resolution:
    The initial state of the sections displayed in the new Administration Group Editor is now consistent with all sections except Properties initially collapsed. As before, excepting the Items section, as you move between groups (or pages of the UI), your configuration of what is expanded and collapsed is maintained throughout the user session.
    (3.7.14, DE467211)
  • Symptom:
    Two jre directories are created during installation of Performance Center.
    Resolution:
    Update the installer sub-module to no longer install the extra JRE.
    (3.7.14, DE467362)
  • Symptom:
    Zooming did not work in Trend Charts (when it was allowed).
    Resolution:
    Zooming now works again in trend charts which support it.
    (3.7.14, DE468020)
  • Symptom:
    When trying to load Odata, it returns "No service was found". The data aggregator logs would indicate exception states "Cannot resolve reference to bean 'entityContainerDef''.
    Resolution:
    Resolved issue that caused the bean to not be loaded.
    (3.7.14, DE468163)
  • Symptom:
    When device is discovered that are related to F5 metric the context tabs for F5, available on Switch Context page, are not available on the Device Context page. When preference is to not associate the device as a switch device, requires manual creating F5 Tabs on Device Context page.
    Resolution:
    When F5 related devices are discovered as generic device and prefer to not associate the device as a switch. Provide an automated script to import context tabs for F5, available on Switch Context page, to the Device Context page.
    (3.7.14, DE468840)
  • Symptom:
    Group Admin Item section could become non responsive after it shows "no items were added to the group" message.
    Resolution:
    Group Admin Item section would  continue to work as expected after it shows "no items were added to the group" message.
    (3.7.14, DE469448)
  • Symptom:
    Manage Groups admin page is missing Device Component Items section.
    Resolution:
    With this fix, Manage Groups admin page properly displays Device Component Items section.
    (3.7.14, DE469491)
  • Symptom:
    When a user enters a bad REGEX into the group rules editor and save/run the rules, the UI appears to hang. A stack trace is displayed in the log.
    Resolution:
    Performance Management
    now handles the exception. If something else throws an exception, it does not hang the UI. The UI now includes REGEX validation to prevent the invalid REGEX in first place.
    (3.7.14, DE469493)
  • Symptom:
    Multi-View Trend for Interface Utilization, Discards and Errors when configured to use bi-directional metrics, In and Out, may render with 'No Data to Display'.
    Resolution:
    Fixed 'No Data to Display' problem with Multi-View Trend for Interface Utilization, Discards and Errors when configured to use bi-directional metrics, In and Out.
    (3.7.14, DE470045, 32099937)
3.7.13 Fixes
  • Symptom:
    Daily scheduled email jobs are sent multiple times during Spring DST transition when email TZ is UTC and scheduled time is between 00:00 - 01:00.
    Resolution:
    Some combinations of Performance Center host TZ (a DST TZ like ESTEDT) and scheduled email TZ (non-DST TZ like UTC), when scheduled time is between midnight and 1AM were leading to incorrect calculation of next run time, causing run time to be repeated until after 1AM. The date math has been corrected to better account for the email timezone when calculating next run time.
    (3.7.13, DE453071)
  • Symptom:
    In the CSV output OpenAPI may not correctly match expand metric family data with it's corresponding configuration entity for example a query like the following devices?$expand=interfaces,portmfs may return metric data in the line for a different interface.
    Resolution:
    With this fix, OpenAPI should match all expand data correctly. (3.7.13, DE455825)
  • Symptom:
    The Viptela VNA plugin is sending data with many disparate sample time values, resulting in many timestamps with few items in the sdn_tunnel_rate table in the DR. This is inefficient for threshold evaluation because each timestamp results in a threshold evaluation query in the DR. This caused long running threshold evaluation times for the NormalizedSDNTunnelInfo metric family and will cause the System Health to indicate a degraded state for Data Aggregator Threshold evaluation.
    Resolution:
    Provided an option in the data collector to modify sample times for each incoming VNA metric so that they appear on a boundary (e.g samples collection at 1:00, 1:02, 1:03, et al ) will have a sample time of 1:00. This greatly reduces the number of timestamps that need to be evaluated for threshold evaluation, thus keeping the evaluation time for the NormalizedSDNTunnelInfo metric family to comparable values as other metric families. (3.7.13, DE459147)
  • Symptom:
    The ACI Console is very unresponsive to mouse clicks in environments that contain a large number of groups.
    Resolution:
    The ACI Console was updated to use more efficient OpenAPI queries for the performance metrics displayed in the dashboard. (3.7.13, DE459510)
  • Symptom:
    When configuring the enhanced Multi-Chart view, multiple metrics support,  when using 'Calculated' axis option the trend line may render off chart.
    Resolution:
    Addressed issue with Multi-Chart view when using 'Calculated' axis option the trend line may render off chart. (3.7.13, DE460982, 31911062)
  • Symptom:
    VM and Host metrics were polled on the wrong values on the vsphere client.
    Resolution:
    VM and Host Metrics are polled now on the correct values from the Vsphere client. (3.7.13, DE461311)
  • Symptom:
    Large XY (trend) charts did not always display in IE 11 without "nudging" the window width.
    Resolution:
    Large XY (trend) charts should now always display in IE 11.
    (3.7.13, DE461321)
  • Symptom:
    The value of Added By column of group items created by a rule could be overwritten if a new item is added to the group manually.
    Resolution:
    With this fix, a new item added to a group manually would not override Added By column value of the items created by a rule.
    (3.7.13, DE463375)
  • Symptom:
    SsoConfig utility does not display French or Japanese text; internal property names are shown instead.
    Resolution:
    Removed misnamed properties files that were hiding English fallback text. Backported translated text from next major release.
    (3.7.13, DE463859)
  • Symptom:
    Memory changes to wildfly standalone.conf do not persist across upgrades. VNA wildfly would revert back to 4GB.
    Resolution:
    VNA upgrade will now persist the VNA_MAX_MEM memory setting in standalone.conf.
    (3.7.13, DE464341)
  • Symptom:
    Some telemetry SQL queries in Performance Center could still run a long time at scale, due to not using available indexes. They could interfere with queries run for UI tasks, making the UI unresponsive.
    Resolution:
    Updated the remaining telemetry SQL queries that needed additional comparisons to cause them to use indexes to speed up queries.
    (3.7.13, DE464384)
  • Symptom:
    Running a Discovery Profile for ranges that were discovered by a VNA Gateway can lead to items being deleted if the IP addresses are pingable but not SNMP contactable.
    Resolution:
    Updates were made to the discovery process which will now prevent items from being deleted when they have previously been discovered by a VNA Gateway.
    (3.7.13, DE466358)
3.7.12 Fixes
  • Symptom:
    Plugin logger exposing passwords in plaintext when category OC_ACQUISITION is set to level DEBUG. Sample message: oc.log:2020-03-11 13:53:43,328 DEBUG (EE-ManagedThreadFactory-default-Thread-3) Viptela_17664d4c-c80b-44b1-8722-aa768696e4e9=Inventory Poll [OC_ACQUISITION] TimReflectionHelper 213  inject vman#2837 to method setPassword.
    Resolution:
    Added logic to the OC_ACQUISITION plugin logger to suppress debugging of password values.
    (3.7.12, DE451580)
  • Symptom:
    Suborgs were not processed earlier in versa plugin.
    Resolution:
    Added support for processing SubOrgs.
    (3.7.12, DE454336, 31812847)
  • Symptom:
    On Deletion of Engine , Few Entities are not deleted .
    Resolution:
    The issue is caused due to inconsistent Hashcode on the Properties. Updated the hashcode.
    (3.7.12, DE458021)
  • Symptom:
    After applying edit changes context page with F5 device there is an error when attempting to save context tab, requiring name entry. Workaround required add entry to context tab prior to being able to save modifications.
    Resolution:
    Address error that required F5 context tab manually add title entry to context tab prior to being to save modifications on context page.
    (3.7.12, DE458272)
  • Symptom:
    After an upgrade of Performance Center, entries for service properties files would add backslash before colons and equals, causing some of the scripts to not run successfully.
    Resolution:
    Updated the upgrade custom code to not add backslashs before colon and equals characters in service properties files.
    (3.7.12, DE459183)
  • Symptom:
    If a Calix chassis has multiple blades that shares the same IP address of the chassis, only one of them can be discovered properly.
    Resolution:
    Improved the Calix blade discovery to properly handle the shared IP on the blades.
    (3.7.12, DE459921)
  • Symptom:
    When not providing a SsoProductCode to the sign-out.jsp, a NullPointerException would be shown.
    Resolution:
    Updated the logout code to return an informative error message, if no SsoProduceCode is provided.
    (3.7.12, DE460006)
  • Symptom:
    Alarms are generated with unknown severity.
    Resolution:
    Made changes in AlarmNotifications to not change the severity to unknown when the alarms are cleared.
    (3.7.12, DE460254)
  • Symptom:
    When global synchronization runs, it may throw a SQL exception like:  com.mysql.jdbc.MysqlDataTruncation: Data truncation: Data too long for column 'SourceIds' at row 672.
    Resolution:
    Fixed the SQL statement so it won't throw a data truncation error causing global sync to fail.
    (3.7.12, DE460948)
  • Symptom:
    The group_group_membership table is not included in the output of the
    caVerticaUtility.sh
    script when exporting customer data repository tables for debugging purposes.
    Resolution:
    The group_group_membership table has been added to the the
    caVerticaUtility.sh
    script export output.
    (3.7.12, DE460967)
  • Symptom:
    When selecting SD-Wan Tunnel/App Path metrics By Component level with the Scorecard Trend may be missing component items.
    Resolution:
    When selecting SD-Wan Tunnel/App Path metrics By Component level with the Scorecard Trend always render component items.
    (3.7.12, DE460970)
3.7.11 Fixes
  • Symptom: Dateparser was not handling few cases of date format during alarm processing in Versa.
    Resolution:
    Added the possible date formats.
    (3.7.11, DE436625)
  • Symptom:
    When an error happens in the middle of a Full Pull data source stage, it can result in all items not sent by the data source to be removed from the Performance Center inventory. This could result in new item ids being assigned to them if they only come from a single data source like Spectrum Global Collections.
    Resolution:
    Updated the Performance Center Full Pull synchronization logic to only mark items, not sent by the data source, after the stage is complete. Example: This should stop Spectrum Global Collections from getting assigned a new Performance Center item id, if there is an error during Full pull for groups.
    (3.7.11, DE437034, 01142425)
  • Symptom:
    Sometimes the Groups page and the Group context picker would not list the groups alphabetically, instead it would display the groups ordered by group id.
    Resolution:
    Groups listed in Group manager and Group context picker will always be sorted by Group name.
    (3.7.11, DE444954)
  • Symptom:
    NFA Stacked Protocol Trend views rendered No Data to Display when time range is greater than 2 hours.
    Resolution:
    Addressed No Data to Display rendered on NFA Stacked Protocol Trend views when time range is greater than 2 hours.
    (3.7.11, DE448349, 31842884, 31828116)
  • Symptom:
    Various security CVEs reported against the version of Consul used by Performance Management.
    Resolution:
    Upgraded the version of Consul to 1.7.2. Please note that for the new version of Consul to be enabled after CA Performance Center is upgraded, the user must run the following commands. If these commands are skipped, the old version of Consul will remain running after the install. Step 1:
    service caPerformance Center-consul stop
    Step 2:
    rm -rf <CA_Performance_Center_Directory>/consul/data/*
    ( e.g. rm -rf /opt/CA/PerformanceCenter/consul/data/* ) Step 3:
    service caPerformance Center-consul start
    .
    (3.7.11, DE448379)
  • Symptom:
    New NFA views on interface context page has default row value of 1000 that creates long running view processing.
    Resolution:
    New NFA views on interface context page modified row items in view selection and default row value to avoid long running view processing, altered views include; IP Performance Protocol- Protocol List, IP Performance ToS - ToS List, IP Performance Host - Hosts List, IP Performance Conversation - Conversation List.
    (3.7.11, DE449080)
  • Symptom:
    When doing a new data aggregator fault tolerant silent install, the data aggregator data directory will use default instead of the DA_DATA_HOME value in properties file provided.
    Resolution:
    Updated silent installer logic to not set the DA_DATA_HOME to a default if already set via properties file.
    (3.7.11, DE449615)
  • Symptom:
    VNA was not sending the correct out of band management IP address for some Versa devices.
    Resolution:
    A new configuration has been added to VNA that enables finding and sending the correct out of band management IP address for Versa devices.
    (3.7.11, DE450038)
  • Symptom:
    When a device has more than 150 interfaces the Alarms View on Alarms tab of device context page fails with error, "An unexpected error has occurred".
    Resolution:
    Addressed issue that Alarms View on Alarms tab of device context page fails with error when a device has more than 150 interfaces.
    (3.7.11, DE450059)
  • Symptom:
    A cross-site scripting vulnerability exists for a url passing javascript to the "pg=" parameter, like this example which raises an alert dialog: https://your_Performance Center_host:8182/Performance Center/desktop/page?timeRange=0&endTime=2020-01-28+14%3A25&startTime=2020-01-28+13%3A25&mn=5&pg='%2Balert(15)%2B'&parentid=13.
    Resolution:
    The vulnerability for the "pg=" parameter has been resolved - the alert in the example url no longer executes.
    (3.7.11, DE451032)
  • Symptom:
    When SNMP MIB table read encountered PARTIAL_FAILURE error during a metric family change detection process, it could create some incomplete component items. Next time when the MIB table is read successfully, the incomplete items will be marked as Not Present. If a device's SNMP agent frequently returns PARTIAL_FAILURE for its MIB table reads, a lot of Not Present components will be created.
    Resolution:
    Improved the metric family discovery not to create any component items when a SNMP PARTIAL_FAILURE is encountered.
    (3.7.11, DE451685)
  • Symptom:
    When LogoutService is called during SAML logout, it throws an exception "net. ibboleth.utilities.java.support.component.UninitializedComponentException: Component has not yet been initialized and cannot be used." while trying to decode the HTTP request.
    Resolution:
    Resolved the exception by initializing the decoder code before using it. This exception was hit due to admins trying to not have Performance Center logout of SAML server when logging out of Performance Center. So this fix also includes code to allow admins to not provide a SingleLogoutService in the metadata XML from the IDP. When Performance Center/sso sees there is no SingleLogoutService, it will just log the user out of Performance Center. The user will still remain logged into SAML.
    (3.7.11, DE452302)
  • Symptom:
    ACI filter Configuration is reset to default after the upgrade and all customer has to re-configure the same.
    Resolution:
    The code fix takes backup of existing filter configurations and retains the same after upgrading VNA.
    (3.7.11, DE452507)
  • Symptom:
    VNA Meraki plugin was fetching the client performance data for only a few of the access points.
    Resolution:
    VNA Meraki plugin now fetches the client performance data for all the access points.
    (3.7.11, DE453478)
  • Symptom:
    The VNA Meraki plugin was not discovering any access point without a lan ip address.
    Resolution:
    The VNA Meraki plugin now discovers access points without a lan ip address.
    (3.7.11, DE453486)
  • Symptom:
    If other (1) and softwareLoopback (24) are re-enabled in Interface or High Speed Interface vendor certifications, users cannot select them in monitoring profile filters for Interface metric family under Type.
    Resolution:
    Updated monitoring profiles Interface metric family filter support for Type to show other and softwareLoopback values, and treat them as supported.
    (3.7.11, DE453520)
  • Symptom:
    Custom Gauge View automatically scales the gauge needle when using the fixed end point combined with percent option.
    Resolution:
    Custom Gauge View allows for option whether to scale the gauge needle when using the fixed end point combined with percent option.
    (3.7.11, DE453522)
  • Symptom:
    Interfaces were not processed when an HTTP Request fails ,which is creating a null or empty file.
    Resolution:
    Added null check for content in the files when HttpRequestFailure occurs.
    (3.7.11, DE454311)
  • Symptom:
    Notification manager could be crashed in a case of large amount of notifications and resulting notifications were not persisted to VNA.
    Resolution:
    The NotificationManager does not specify any LockType implicitly which means it is using LockType.WRITE which is the default behavior of Singleton EJBs. But the NotificationManager was designed to be thread-safe and this LockType.WRITE is an overhead and can be replaced with LockType.READ.
    (3.7.11, DE454535)
  • Symptom:
    After upgrading to 3.7.10 multiple symptoms were seen on certain data aggregators that were monitoring SDWAN environments with VNA. 1. High CPU Usage after upgrade caused Performance Center to lose contact with the data aggregator (see: https://knowledge.broadcom.com/external/article?articleId=188059). 2. Changes in the monitored VNA environment were not being reflected in CAPerformance Center. These include, but may not be limited to: Sites, Tunnels and SLAPaths. 3. Prior to and after upgrading to 3.7.10 "duplicate" / "stale" items and SDN Domains were seen in Performance Center under the SDN Domains group.
    Resolution:
    The data aggregator's processing of VNA data has been updated to be more efficient during the processing of data from SDWAN environments. The changes include:1. A reduction in the upper limit of the number threads available for SDN processing to 1/3 of the CPUs on the data aggregator. 2. General improvements to the VNA processing to improve the efficiency of processing VNA updates from SDWAN environments. 3. Additional protections have been added to the VNA processing to resolve missing items seen after upgrading. 4. SDN processing in the data aggregator has been updated to prevent Items from being created for SDN Gateways that had already been deleted from the DA.
    (3.7.11, DE455285)
  • Symptom:
    When using VNA as an aggregator inventory items/keys are removed after upgrade. This is especially applicable to Spectrum. After upgrade it was unable to reconcile the changes which has the effect of the system creating duplicate items.
    Resolution:
    Migration script modified to only modify ID cache if VNA is not operating in "aggregator" mode.
    (3.7.11, DE456076)
  • Symptom:
    For consolidated devices, Performance Center maintained IP address of the device that was sync'ed up first (data aggregator device IP was not given priority over spectrum).
    Resolution:
    Updated Global sync to give data aggregator device IP priority over Spectrum device IP during device consolidation based on primary-secondary IP matching.
    (3.7.11, DE456272)
  • Symptom:
    Reporting views using business hours, that have multiple timezones to consider, can take minutes to run the query in the database. This is partly due to the amount of memory the database reserves to run the query, which is based on 65k characters for timezone values.
    Resolution:
    Updated the sql view, for comparing timezone, to select only 255 characters from the timezone attribute. This results in far less memory being needed to run the query. It could result in steps of the query now being multi-threaded, where before it could not due to memory requirements. This could result in some business hour queries running much faster than before.
    (3.7.11, DE456460)
  • Symptom:
    Web service calls to Spectrum for Alarm Console or the subscription alarm service may timeout after 100 seconds. This usually is due to Spectrum's response time to the request being made.
    Resolution:
    Updated the timeout to 120 secs, but also made it configurable via the netqosportal.general table or Performance Center admin debug global attributes page. To increase the Alarm Console timeout, run: REPLACE INTO netqosportal.general VALUES('AlarmView_WebService_Timeout', '<timeout in seconds>');  To increase the Spectrum alarm status subscription timeout, run: REPLACE INTO netqosportal.general VALUES('SpectrumStatus_WebService_Timeout', '<timeout in seconds>');.
    (3.7.11, DE456697, 20294595)
  • Symptom:
     Performance Center global synchronization fails if duplicate tunnels and/or SLA paths are synchronized from any data source. Error similar to the following would be found in
    DMservice.log
    file upon failure: "MySQLIntegrityConstraintViolationException: Duplicate entry '5868174-14' for key 'PRIMARY".
    Resolution:
    Updated the global sync SQL calls to not error when duplicate tunnels/SLA paths are sent from data source(s).
    (3.7.11, DE456711)
  • Symptom:
    Data Aggregator operations that require internal attribute reads such as data aggregator REST calls, VNA inventory processing, etc are subject to slowness at high concurrency due to excessive use of java locking objects.
    Resolution:
    Refactored Item Repository read functionality to more efficiently use java locking objects.
    (3.7.11, DE456931)
  • Symptom:
    After VNA upgrade to 3.7.10, Viptela polling for data could fail with error in the
    gateway.log
    file for:
    groovy.lang.MissingPropertyException: No such property: outFile for class: ViptelaHttpFunction
    . This is caused by an extra parameter in a function call not being needed. Only seen when there is more than one page of data read during polling phase for an data type.
    Resolution:
    Resolved the function call by removing the extra parameter.
    (3.7.11, DE457801)
3.7.10 Fixes
  • Symptom:
    1. The data aggregator falls behind in processing VNA changes. This results in models missing from existing VNA Domains or else VNA Domains not being created after configuring an additional SDN Gateway.2. The groups under a VNA Domain disappear after toggling the Administration Status of the applicable SDN Gateway.
    Resolution:
    Enhancements were made to the VNA Integration to allow Inventory updates to be processed much faster. These enhancements also added more robust logic to avoid exceptions previously seen during response processing and also to prevent Items and Relationships from incorrectly being deleted when toggling the Administration Status of a VNA.
    (3.7.10, DE439424)
  • Symptom:
    Running the
    testssl.sh
    script against Performance Center shows vulnerabilities for
    Secure Client-Initiated Renegotiation
    ,
    LOGJAM (CVE-2015-4000)
    , and
    LUCKY13 (CVE-2013-0169)
    .
    Resolution:
    Vulnerabilities for
    Secure Client-Initiated Renegotiation
    ,
    LOGJAM (CVE-2015-4000)
    , and
    LUCKY13 (CVE-2013-0169)
    have been resolved in Performance Center, and are no longer flagged by the
    testssl.sh
    script.
    (3.7.10, DE440319)
  • Symptom:
    OpenAPI could throw a DB error when querying metric family entity if the corresponding metric family has been modified in Performance Center, for example updating Max and percentile.
    Resolution:
    With this fix, OpenAPI should truck all changes in the DB schema and return requested metric family data without error.
    (3.7.10, DE448656, 20312928, 31812895)
  • Symptom:
    After configuring Scorecard Trend view with baseline metric renders 'No Data to Display'.
    Resolution:
    Addressed issue when selecting baseline metric with Scorecard Trend view renders 'No Data to Display'.
    (3.7.10, DE449375)
  • Symptom:
    sdn_poll_timeout_minutes parameter was not getting added in /etc/VNA.cfg file in upgrade case, hence warn logs were filing up in gateway.log file.
    Resolution:
    Handled in upgrade case so that sdn_poll_timeout_minutes parameter shall added /etc/VNA.cfg file. Updated ApPerformance Centeronfig.java file to read the value only once.
    (3.7.10, DE450952)
  • Symptom:
    When sending email via STARTTLS, Performance Center would only allow TLSv1.0. If email server was TLSv1.1 or TLSv1.2, the email server would reject the email connection.
    Resolution:
    Updated Performance Center to support TLSv1.2, TLSv1.1 and TLSv1.0 when email server supports STARTTLS.
    (3.7.10, DE451724)
3.7.9 Fixes
  • Symptom:
    Configuring an On-Demand Report or Dynamic Trend View, if setting context to be fixed and when number of items exceeds 5000 the warning status message hides the add button on the form.
    Resolution:
    Fixed warning status message when hides the add button while configuring an On-Demand Report or Dynamic Trend View to use fixed context if number of items exceeds 5000.
    (3.7.9, DE432487)
  • Symptom:
    The adminStatus and operationalStatus of the OpenDayLight controller in VNA were not captured correctly.
    Resolution:
    The adminStatus and operationalStatus of the OpenDayLight controller in VNA are now captured correctly.
    (3.7.9, DE432780)
  • Symptom:
    One of out services displays a stack trace to the client when an exception is thrown. This potentially gives and attacker information about the underlying code.
    Resolution:
    The stacktrace is now logged and a standard server error code is sent back.
    (3.7.9, DE438461)
  • Symptom:
    The global search could hang on a search string that contains the '=' character.
    Resolution:
    With this fix, the global search now functions properly. If the search string contains the '=' character, add a new '=' character to the beginning of the search string.
    (3.7.9, DE442778, 20140853)
  • Symptom:
    In the CA-Blue and CA-Gray themes, drop-down menus associated with the top navigation menu buttons (for example, Dashboards, Inventory) would obscure the button used to open them if the content of the drop-down menu was larger than the vertical space between the button and the bottom of the browser window.
    Resolution:
    In the CA-Blue and CA-Gray themes, drop-down menus associated with the top navigation menu buttons (e.g., Dashboards, Inventory) always are displayed below the buttons used to open them.
    (3.7.9, DE444091)
  • Symptom:
    Reports using custom group of SDWAN tunnels will show no data. Custom groups of devices will show data for SDWAN tunnels they are a participant in.
    Resolution:
    Updated data aggregator SQL view to check if the SDWAN tunnel is in the group, in addition to either device of the tunnel.
    (3.7.9, DE444464)
  • Symptom:
    After configuring scheduling CSV report with the All Pages option email content and Manage Schedule Reports contain hyperlink instead of an attachment.
    Resolution:
    Addressed issue when scheduling CSV generated report with the All Pages option email content and Manage Schedule Reports contain hyperlink, should be handled as an attachment.
    (3.7.9, DE444927)
  • Symptom:
    When Performance Center restarts, if MySql does not start in 6 seconds, then the Performance Center services fails to start.
    Resolution:
    Fixed comparison in MySql check script to do integer comparison, not string compare that resulted in the check exiting earlier than expected. Also increased check to 120 seconds.
    (3.7.9, DE445174)
  • Symptom:
    NFA Hierarchy sub-table for By Host and By Protocol will intermittently not display the interface data, as item are selected in top table the lower table is still blank.
    Resolution:
    Addressed intermittent issue when selecting item in top table the lower table does not render interface data with NFA Hierarchy sub-table for By Host and By Protocol view.
    (3.7.9, DE446176)
  • Symptom:
    Vcenter is sending IP address in invalid format. Which is causing Persistence failures.
    Resolution:
    Added check for Invalid IP addresses coming from the Vcenter . Plugin will lookup for IP address for the VM , if the ip address is invalid.
    (3.7.9, DE446231)
  • Symptom:
    During global synchronization stage on Performance Center, a SQL exception may be thrown during tenant or IP domain stage about truncating LocalID in ds_items table. Large model handles from Spectrum were causing the SQL exception.
    Resolution:
    Updated the SQL call to do string comparison instead of numeric. The LocalID column is a string and should've been comparing as a string but was using numeric comparison that could result in truncation issues.
    (3.7.9, DE447006)
  • Symptom:
    After configuring scheduling CSV report with All Pages option email content and Manage Schedule Reports contain hyperlink instead of an attachment.
    Resolution:
    Addressed issue when scheduling CSV generated report with All Pages option email content and Manage Schedule Reports contain hyperlink, should be handled as an attachment.
    (3.7.9, DE447159)
  • Symptom:
    When selecting SD-Wan Tunnel/App Path metrics By Component level with the Scorecard Trend may be missing component items.
    Resolution:
    When selecting SD-Wan Tunnel/App Path metrics By Component level with the Scorecard Trend always render component items.
    (3.7.9, DE448032)
  • Symptom:
    Installers would complain about SLES 12 SP4+ do not pass kernel check.
    Resolution:
    Updated installers to pass kernel check with SLES 12 SP4, and SP5.
    (3.7.9, DE449181, 20310176)
3.7.8 Fixes
  • Symptom:
    There might be cases where Event Manager does not automatically re-register with data aggregator to receive events.
    Resolution:
    Logging has been added to help capture occurrences.
    (3.7.8, DE391457, 01218626)
  • Symptom:
    The time displayed in the time picker (upper right corner of dashboard) shows incorrect/next year for the last week of the year. December 30, 2019 Displays as December 30, 2020.
    Resolution:
    Modified the date formatter to use yyyy instead of YYYY which formats the last week of the year correctly. The time displayed in the time picker will show the correct year for the last week of the year ie. December 30, 2019.
    (3.7.8, DE433614, 20197781)
  • Symptom:
    Engine which is deleted and not cleaned up properly is remaining in the collector folder. On WildFly restart, VNA is considering it a running engine and trying to start the same.
    Resolution:
    On Restart of VNA, now it will check if the engines exist in the DB. All engines which do not exist are deleted.
    (3.7.8, DE437603)
  • Symptom:
    One of out services would display a stack trace to the client when an exception was thrown. This potentially gives and attacker information about the underlying code.
    Resolution:
    We logged the stacktrace and sent back a standard server error code.
    (3.7.8, DE438461)
  • Symptom:
    Time bar view does not sort as expected when initially rendered on the page, required column selection to sort correctly.
    Resolution:
    Addressed issue where time bar view does not sort as expected when initially rendered on the page, required column selection to sort correctly.
    (3.7.8, DE440194)
  • Symptom:
    The data aggregator misinterprets an exception from the database as an outage and shuts down.
    Resolution:
    Updated the regular expression for expected exceptions from the database that should not be interpreted as an outage.
    (3.7.8, DE440348)
  • Symptom:
    Cannot delete group using Performance Center REST service, when the user has My Assigned Groups.
    Resolution:
    Fixed permission checking code in Performance Center REST API.
    (3.7.8, DE440557)
  • Symptom:
    CSV report export only includes first page for legacy systems that upgraded to Service Pack 3.7.4 or above.
    Resolution:
    Addressed upgrade Service Pack 3.7.4 or above issue that CSV report export only includes first page.
    (3.7.8, DE441538)
  • Symptom:
    When Domain is a selected inventory column rendering No Data when applying filter to table view search.
    Resolution:
    Addressed issue when applying filtered search to table with Domain inventory column that sometimes renders No Data to Display.
    (3.7.8, DE441661)
  • Symptom:
    After configuring a Scorecard table that contains percentile metric; when removing percentile metric from list metrics with hierarchy option still renders percentile entry.
    Resolution:
    Addressed issue when removing percentile metric from Scorecard Table with hierarchy option still renders percentile entry in view.
    (3.7.8, DE441793)
  • Symptom:
    Certain VNA entities are deleted and recreated every poll cycle even though there are no changes occurring in the environment.
    Resolution:
    Updated VNA Persistence logic in VNA to protect against the mentioned scenario. With the new implementation if keys are to be removed from the ID service, we first check to be sure that the key is associated with the Entity's ID. If it is associated with a different ID, we will print a warning and no longer remove it from the ID service. This will prevent the below from happening.
    (3.7.8, DE442571)
  • Symptom:
    SslConfig on Performance Center and the
    sslConfig.sh
    script on data aggregator would use SHA1WithRSA when generating private key for HTTPS use.
    Resolution:
    Updated both tools to use SHA256WithRSA when generating private key for HTTPS.
    (3.7.8, DE442855)
  • Symptom:
    When Spectrum system time is behind Event Manager system time, events may not be getting created in Spectrum when pulled from Event Manager.
    Resolution:
    Updated Event Manager to only handle when current time on Event Manager is older than the end timestamp being requested by Spectrum. It will time shift the start/end request timestamps to EM local timestamp. EM will no longer time shift forward Spectrum start/end request timestamps, if older than EM local timestamp.
    (3.7.8, DE443851)
  • Symptom:
    When business hour filter is applied to context page incorrectly shown on view subtitle for data source not related to data aggregator.
    Resolution:
    Addressed view subtitle showing business hour filter on context page for data sources that are not related to data aggregator.
    (3.7.8, DE444080)
  • Symptom:
    Generate URL not always initially rendering correctly when when view container is enabled on URL Generate dialog.
    Resolution:
    Addressed problem when generate URL that view is collapsed if view container is enabled on URL Generate dialog.
    (3.7.8, DE444200)
  • Symptom:
    If the user's time zone caused "last 8 hours" to shift into the prior day, then the dashboard's time display incorrectly says Today instead of yesterday's date.
    Resolution:
    The fix was to include the timezone when checking to see if the date is today's date.
    (3.7.8, DE444215)
3.7.7 Fixes
  • Symptom:
    When the Data Collector runs as a non-root user, the user would be unable to re-assign the data collector to a different tenant and/or IP domain due to permissions to restart the data collector from inside the data collector.
    Resolution:
    Updated the restart logic to call sudo to restart the data collector if it is running as non-root user. For more information about the additional sudo command to allow the data collector to run as the install user, see the documentation.
    (3.7.7, DE418500, 01363625)
  • Symptom:
    Geo location attributes cleared in Performance Center group editor for the Site groups would not be cleared on DA, so the OpenAPI will show the last numerical value before they were changed to blank.
    Resolution:
    With this fix, the geo location attributes cleared in Performance Center group editor for the Site groups would be cleared on data aggregator and OpenAPI provided values would match Performance Center values.
    (3.7.7, DE420818)
  • Symptom:
    Trend/Table views with Business Hours applied shows no shading in Trend for time ranges longer than 14 days.
    Resolution:
    The Trend in Trend/Table view has been corrected to request hourly data instead of daily for longer time ranges when Business Hours filtering is applied. This allows the shading to be rendered correctly.
    (3.7.7, DE429361)
  • Symptom:
    Emails could not be sent securely if email server supported STARTTLS.
    Resolution:
    Updated Performance Center email properties to enable STARTTLS on all emails being sent. Removed Use SSL button, as Performance Center will try and send all emails securely if email server supports STARTTLS.
    This now requires the email server certificate and any intermediate/root certificates to be imported into
    /opt/CA/jre/lib/security/cacerts
    , so the email server certificate will be trusted on connection. For more information, see Set the Email Server.
    (3.7.7, DE432418)
  • Symptom:
    Vendor cert "Cisco IPSLA Ethernet DMM and IP Statistics" and "Cisco IPSLA Ethernet SLM (Synthetic Loss Measurement) Statistics" were coupled with vendor cert "Cisco IPSLA Ethernet Jitter Precision Statistics". If "Cisco IPSLA Ethernet Jitter Precision Statistics" is not supported, the other two vendor certs would not be discovered as supported even though they were configured on the devices.
    Resolution:
    Decoupled the three vendor certs so that they can be discovered properly and independently.
    (3.7.7, DE433784)
  • Symptom:
    Metric family "Response Path Test Ethernet Jitter" missed two metrics: FrameLossRatioSrcDest and FrameLossRatioDestSrc.
    Resolution:
    Added two metrics FrameLossRatioSrcDest and FrameLossRatioDestSrc to metric family "Response Path Test Ethernet Jitter".
    (3.7.7, DE433788)
  • Symptom:
    When installing the consul service for Fault Tolerance in the proxy and data aggregator, the installer is not asking which IP address to bind to, if there are multiple IPs on the machine.
    Resolution:
    Updated the proxy and data aggregator installers to prompt for which IP address to bind consul service to, if there are multiple IPs on the machine.
    (3.7.7, DE434128)
  • Symptom:
    When rendering baseline and core metrics on same trend chart when resolution is less than an hour trend line for baseline metric appears as dots that are not connected due to gap data points.
    Resolution:
    Addressed issue with baseline metric appearing as dots, not connected when rendering baseline and core metrics on same trend chart when resolution being applied is less than an hour.
    (3.7.7, DE434424)
  • Symptom:
    Percentage of Actively Memory used was not calculated for ESX hosts.
    Resolution:
    Updated the vendor cert to support the percentage of Active memory used.
    (3.7.7, DE435721)
  • Symptom:
    Group Scorecards when combining total rate and percentiles metrics throws query processing error.
    Resolution:
    Addressed error when combining total rate and percentile metrics on the Group Scorecards view.
    (3.7.7, DE435901)
  • Symptom:
    If you select an Interface in Inventory, and then select the IP Performance Tab, and then zoom into one of the charts. After clicking Apply To Dashboard for that chart, the time does not change in the Time Display for the dashboard.
    Resolution:
    The dashboard's time range field was updated to reflect the "Apply to Zoom" feature.
    (3.7.7, DE436188)
  • Symptom:
    Performance poll was running sequentially for various device types , which is consuming time thereby missing every other poll.
    Resolution:
    Added Multi Threading for processing performance poll.
    (3.7.7, DE436698)
  • Symptom:
    There is a spike while calculating Delta for VNA metrics on the reports when the engine or VNA is down for a particular time.
    Resolution:
    Data collector will now discard first poll after the VNA or Engine is down, if the value is more than 2 polls.
    (3.7.7, DE436702)
  • Symptom:
    NetOps user interface experiencing performance issue when attempting to render Trend Chart with Events reporting when greater than monthly period.
    Resolution:
    Added safeguard to limit the time range to 31 days when attempting to render Trend Chart with Events reporting to address performance impact./\n .
    (3.7.7, DE436873)
  • Symptom:
    Unable to Print "SDN/NFV vSwitch Performance Overview" page to PDF, yields "Property missing: Title" exception in log.
    Resolution:
    This could occur on systems that were upgraded from a pre-3.0.0 release due to a view removed from this page in 3.0.0. The properties of the deprecated view are now restored and the page can now be printed or edited without "Property Missing" exception.
    (3.7.7, DE436879)
  • Symptom:
    After upgrade of a very high scale system, Performance Center user REST services are significantly slower.
    Resolution:
    Optimized telemetry queries.
    (3.7.7, DE437335, 20103983)
  • Symptom:
    When a data aggregator thought data collectors were disconnected, it aborted all currently running discoveries. Due to a un-handled exception in the aborting process, the counter of the running discoveries was not decreased. When the number of running discoveries reached to the maximum number of concurrent discoveries, all discoveries will be stuck on Performance Center.
    Resolution:
    Improved the discovery abort process and handled all the exceptions so that the discovery counter can work properly.
    (3.7.7, DE437748)
  • Symptom:
    Distributed Item Repository communication between data aggregator and all data collectors can be disrupted if one data collector gets into a bad state. An example is seen when a data collector broker stops consuming messages from the DA, causing a back up in its DIM.request queue. Once the data aggregator broker detects this queue is 100% full, all other DA/data collector communication is disrupted.
    Resolution:
    Reworked data aggregator side Distributed Item Repository processing to isolate workloads per Data Collector so that when one Data Collector gets into a bad state, communication is not disrupted for all Data Collectors.
    (3.7.7, DE437782)
  • Symptom:
    With On-Demand report view option Chart per Item with Multiple Metrics when reporting at the component level item titles on chart can mismatched with graph data.
    Resolution:
    Addressed titles on chart mismatched with graph data with On-Demand report view option Chart per Item with Multiple Metrics when reporting at the component level.
    (3.7.7, DE437961, 20108850)
  • Symptom:
    When the page context for the Alarm Console is set to a large NetOPs group, Spectrum may return an error due to a limitation in the Spectrum rest api.
    Resolution:
    Modified the way that NetOPs groups are passed as a filter to Spectrum.
    (3.7.7, DE438103)
  • Symptom:
    Data Collector could fail to start due to a NULL pointer exception when it tries to sync items from the data aggregator.
    Resolution:
    Improved the sync process between data aggregator and data collector to handle all possible exceptions properly.
    (3.7.7, DE438315)
  • Symptom:
    When VNA receives a time filter larger than 24hrs, it throws an error, which results in client not able to get further updates.
    Resolution:
    VNA will send a FULL update on receiving a time filter larger than 24hrs, which will help auto renew connections with the client.
    (3.7.7, DE438556)
  • Symptom:
    In a system with VNA installed, the DA/Performance Center sync process could be very expensive. It would check the item creation time on all items in the system to find out the newly created VNA domain group items.
    Resolution:
    Improved performance of DA/Performance Center sync process by dramatically narrowing down the search scope.
    (3.7.7, DE438808)
  • Symptom:
    Tunnels from the router(customer gateay) which was discovered through snmp are missing./\nIts not the issue of snmp reconciliation.
    Resolution:
    There was an aws technology specific code written in data aggregator based on the vendor attribute, due to which further code that builds tunnels getting restricted. Upon discussion with Abe/Li this code cleaned-up, that fixed the issue.
    (3.7.7, DE438825)
  • Symptom:
    If any ec2 is in terminated state in AWS console and its ip's are deleted and empty/null ip not handled, due to which entire inventory failed.
    Resolution:
    Terminated ec2's are handled so that upon null/empty ip, the entire inventory poll shall not fail.
    (3.7.7, DE439129)
  • Symptom:
    When rendering shade area on trend chart related to standard deviation the processing performance impacted memory resource usage.
    Resolution:
    Address the processing performance impact when rendering shade area on trend chart related to standard deviation.
    (3.7.7, DE439284)
  • Symptom:
    Report Analyzer view, Interfaces over Threshold, renders missing image icon on interface name column with Firefox.
    Resolution:
    Addressed the missing image icon with Firefox appearing on interface name column when rendering Report Analyzer view, Interfaces over Threshold.
    (3.7.7, DE440550)
  • Symptom:
    When upgrading Performance Center, the service (Performance Center, DM, EM, sso) properties files, that store settings, would lose any user added entries.
    Resolution:
    Modified install to preserve any user added settings in the service properties files.
    (3.7.7, DE440676)
3.7.6 Fixes
  • Symptom:
    Under specific conditions some newly created Performance Center items at the end of current sync cycle like tenants, groups, domains may not be synced to the data source. Workaround is to perform a full sync.
    Resolution:
    With the fix, newly created Performance Center items at the end of current sync cycle would be synced to the data source on the next sync cycle.
    (3.7.6, DE348850, 01352680)
  • Symptom:
    At times clicking on group in the group tree on Manage Groups page takes a large amounts of time and make the group is not accessible.
    Resolution:
    With the fix clicking on group in the group tree on Manage Groups page should take a reasonable amount of time.
    (3.7.6, DE415423, 01345758)
  • Symptom:
    On Migration from Wildfly 10 to Wildfly 13 & 17, the JBOSS_CONSOLE_LOG property changed in the wildfly initialization Script.
    Resolution:
    Added the JBOSS_CONSOLE_LOG to /dev/null in Wildfly 13 and 17.
    (3.7.6, DE420361)
  • Symptom:
    When group memberships in Performance Center are removed during global synchronization (and moved to deleted table), due to item being deleted, it would not remove the BY_RULE flag from the relationship. This caused the relationship to appear in Excluded Items.
    Resolution:
    When removing a group relationship and the relationship is moved to the deleted table, we will remove the BY_RULE flag. Now the relationship won't be considered for Excluded Items list, and if the item comes back, group rules can re-add the relationship.
    (3.7.6, DE422002, 01287437)
  • Symptom:
    "Scheduled Reports" List and Edit dialog intermittently show incorrect time for the schedule run, and are off by the number of hours between the user's time zone and the system time zone. If saved while in this state, the scheduled report now runs at the incorrectly offset time. If not saved, the scheduled email still runs at the originally defined time.
    Resolution:
    Certain dashboard views were modifying the thread's TimeZone to the user's TZ and not restoring it, which injected an incorrect time zone offset in the Scheduled Reports UI. This has been corrected.
    (3.7.6, DE423920)
  • Symptom:
    Tenant themes may switch back to CA-Blue after changing them to another theme. This is because the data aggregator tenant entry in Performance Center data source tenant table still had CA-Blue.
    Resolution:
    Updated data aggregator pull sync code to send up tenant and ip domain changes sent during previous push sync, so data aggregator data source table entries are updated with new theme.
    (3.7.6, DE424092)
  • Symptom:
    When using the Performance Center REST service to clone a user who uses external (LDAP) authentication, the new user cannot log in.
    Resolution:
    For users with external authentication, modified the clone operation to set the correct authentication flags. For users with internal authentication, modified the clone operation to expire the password immediately; new users will have to change their password on login. .
    (3.7.6, DE425283)
  • Symptom:
    Inventory not detected by Spectrum in case of Aggregator scenario.
    Resolution:
    We Improved the logging to stop such issues in the future.
    (3.7.6, DE427652)
  • Symptom:
    After a VNA gateway has been removed from the data aggregator, temporary VNA items and relationships are not cleaned up.
    Resolution:
    Update the delete VNA Gateway logic to also clean up temporary SDN items.
    (3.7.6, DE427863)
  • Symptom:
    Component name aliases in data aggregator were being synced to Performance Center, and if longer than 255 characters, would throw SQL error about data truncation. This caused data aggregator PULL phase of synchronization to fail.
    Resolution:
    Updated data aggregator synchronization code to no longer send component name aliases to Performance Center. data aggregator should never have been synced to Performance Center. Performance Center should only be pushing the name alias to DA.
    (3.7.6, DE428603)
  • Symptom:
    VNA device items were not showing up within the VNA group hierarchy.
    Resolution:
    Updated the data collector to correctly handle situations where VNA sends more than one type of relation for an entity.
    (3.7.6, DE430894)
  • Symptom:
    When using the new vendor certification UI and trying to select an attribute/OID to apply to an expression, it throws an error about wizard,button.moveTop missing in the performance center
    service.log
    file, and not show any attributes to select.
    Resolution:
    Fixed the UI so it displays the
    variables
    table and buttons correctly.
    (3.7.6, DE431554)
  • Symptom:
    Inventory view for Virtual Interfaces was incorrectly labeled as an Event List view. The incorrectly label view confused user when configuring dashboard page.
    Resolution:
    Correct the name used for a Virtual Interfaces inventory view that was labeled as an Event List view. .
    (3.7.6, DE431708)
  • Symptom:
    VNA is not doing a refresh for all the DVS Ports.
    Resolution:
    Added a refresh of DVS Port state for the missing ports.
    (3.7.6, DE431723)
  • Symptom:
    After doing a global search, the action buttons (On Demand, Update Polling and Manage Life Cycle) were missing from the search results grid.
    Resolution:
    Un-hid action buttons and changed dialogs to nested dialog model id and added an extra close to life cycle, on demand and update polling dialogs.
    (3.7.6, DE433440)
  • Symptom:
    Installation of VNA is failing when customer has umask other than 0022.
    Resolution:
    Updated documentation to state that permission to the folder must be setup so installer can access it.
    (3.7.6, DE433484)
  • Symptom:
    A recurring message of dropped relations appears in the data aggregator whenever it receives an inventory update from the VNA.
    Resolution:
    Change the logging level of the message to debug so we aren't always seeing the message, just for debugging.
    (3.7.6, DE434146)
  • Symptom:
    When Spectrum pulls events from Event Manager, there was a possibility to see duplicate event. This is due to the item being contributed by multiple data sources, and not using DISTINCT in EM DB queries.
    Resolution:
    Updated the EM DB query to use DISTINCT when getting events to send to Spectrum.
    (3.7.6, DE434444)
  • Symptom:
    The DataAggregator may fail to start due to an ArrayIndexOutOfBoundsException when initializing the RelationshipImporter. This is due to inconsistent values in the attributes used to persist previously imported relationships. The inconsistency could have been caused by a bug in the item repository persistence layer where attribute writes that should have been committed in a single transaction were committed individually.
    Resolution:
    Fixed a bug in the item repository persistence layer's transaction management. Fixed the relationship importer's initialization process to re-initialize the history of previously imported relationships if the history attributes are inconsistent. If the attributes are found to be inconsistent a error is logged and any automatically generated relationships that had previously been removed will return and need to be removed again via the UI.
    (3.7.6, DE434597)
  • Symptom:
    After installing Performance Center database on a new machine during a migration, Performance Center upgrade will not ask for the new machine name and credentials. It will assume the properties files contain the correct DB connection credentials.
    Resolution:
    Updated Performance Center installer during Services upgrade, so if the properties file contains localhost, it will prompt the user for the new hostname of Performance Center DB host.
    (3.7.6, DE434663)
  • Symptom:
    In a multi-datasource environment, tenants and IP domains are deleted for all data sources during sync if one of the data sources indicates it does not have those tenants or domains.
    Resolution:
    Update sync to account for multiple data sources being attached to domains and tenants.
    (3.7.6, DE434930)
  • Symptom:
    Items are missing from VNA due to persistence failures while processing messages.
    Resolution:
    Updated the VNA persistence error handling to prevent failures from rolling back message processing. Additionally, updated the persistence logic to correctly persist additional objects that were previously leading to persistence failures.
    (3.7.6, DE435483, 20099081,20093248)
  • Symptom:
    After upgrade to 3.7.5, the Dashboards menu may contain duplicate entries. Upgrade was calling create_dashboards, when it should only be calling create_new_dashboards.
    Resolution:
    Updated installer to not call create_dashboards on upgrade.
    (3.7.6, DE435768)
  • Symptom:
    When I set the NetOPs group context to be a Spectrum Global Collection, I only see device level alarms, and not interface level alarms. One Click shows both.
    Resolution:
    Modified the way that the global collection filter is sent to Spectrum, so that the alarm view will now show device and interface level alarms when the alarm view context is set to a synchronized spectrum global collection.
    (3.7.6, DE436170)
  • Symptom:
    Items are missing from VNA due to persistence failures while processing messages.
    Resolution:
    Updated the VNA persistence error handling to prevent failures from rolling back message processing. Additionally, updated the persistence logic to correctly persist additional objects that were previously leading to persistence failures.
    (3.7.6, DE436747)
  • Symptom:
    OK and Cancel buttons on On Demand's Add/Remove Item/Groups dialogs were missing the text (OK, Cancel). They were blank buttons. Also, Add/Remove Group for the On Demand from Search Results emitted a 500 Server error.
    Resolution:
    Modified the properties files to reflect new page (18 for nested dialog). also changed add/remove groups to nested dialog.
    (3.7.6, DE437351)
3.7.5 Fixes
  • Symptom:
    If the ActiveMQ process on the data aggregator or data collector runs Out Of Memory it will not be restarted.
    Resolution:
    Added '-XX:OnOutOfMemoryError=/opt/IMDataAggregator/scripts/activemq stop' as JVM parameter to data aggregator and data collector broker processes.
    (3.7.5, DE409714, 01294980)
  • Symptom:
    Custom PE Interfaces context page suppressed CBQOS Context Tab after upgrading to 3.6 service pack.
    Resolution:
    Resolved the Custom PE Interfaces context page being suppressed CBQOS Context Tab after upgrading to 3.6 service pack.
    (3.7.5, DE414383, 01317817)
  • Symptom:
     Interface context tab IP Performance with NFA view Stacked Tos Trend chart displayed "No Data To Display".
    Resolution:
    Corrected configuration issue that lead to Interface context tab IP Performance with NFA view Stacked Tos Trend chart displaying "No Data To Display".
    (3.7.5, DE417615)
  • Symptom:
    RIB queries that target a Data Aggregator and do not include any GROUP BY, ORDER BY, or LIMIT clause started to fail due to parsing errors on the Data Aggregator after a recent Data Aggregator upgrade.
    Resolution:
    Fixed the Data Aggregator's RIB query parsing logic for this scenario. .
    (3.7.5, DE418891, 01367393)
  • Symptom:
    When the system Performance Center is set to have localization other than en_US.UTF-8, the caperfcenter_console service fails to start.
    Resolution:
    In the caperfcenter_console script, locally set the localization to en_US.UTF-8. This allows the expected error checking to proceed and let the process start up.
    (3.7.5, DE418968, 01366004)
  • Symptom:
    For Viptela environments with a large number of devices/tunnels the performance polling will result is high memory usage on the VNA. This will cause the VNA to run out of memory and/or slow down.
    Resolution:
    Updated Viptela plugin polling logic to more efficiently process lists of polled data.
    (3.7.5, DE420663)
  • Symptom:
    When using the VNA VMware plugin and the Data Aggregator is on a virtual machine managed by the plugin, removing the VNA gateway could cause the data aggregator item to be deleted and re-created on startup.
    Resolution:
    Added check to the VNA gateway removal code to not remove any data aggregator or data collector device items, and just remove the VNA facets off them instead.
    (3.7.5, DE422039)
  • Symptom:
    After a long period of operation, MySQL process eventually consumes a large percentage of overall system memory even though the process memory settings are properly sized. This is due to MySQL's performance_schema, which is an in-memory schema and not used for Performance Center, being enabled by default.
    Resolution:
    The MySQL performance_schema has been disabled in the my.cnf file installed with Performance Center.
    (3.7.5, DE422458)
  • Symptom:
    In a multiple data sources environment, any device changes from other data source could trigger Data Aggregator to run the inventory discovery during the Performance Center and data aggregator sync process.
    Resolution:
    Improved the sync process so that only if a new device IP is pushed down from Performance Center can the inventory discovery be run in Data Aggregator.
    (3.7.5, DE423494)
  • Symptom:
    When re-connecting to a VNA Gateway, the data aggregator will delete relationships that should still exist. This ultimately leads to group memberships disappearing in the VNA Domains hierarchy in Performance Center.
    Resolution:
    Updated the data aggregator logic around determining which relations should be removed when re-connecting to a VNA Gateway. data aggregator will no longer delete relationships that should still exist.
    (3.7.5, DE423878)
  • Symptom:
    VNA Notifications are dropped if the value of the string is more than the mysql database char size.
    Resolution:
    The Description field and the cause field will be auto -truncated if the size of the chars is more than 1024 and 255 chars respectively.
    (3.7.5, DE428342)
  • Symptom:
    Performance Center's console service is unable to start on a system that does not have wget installed.
    Resolution:
    In the Performance Center console start-up script, check if wget is installed and report an error to the user if it is not.
    (3.7.5, DE428440)
  • Symptom:
    The ipadress from Vcenter is sent with invalid spaces at the end of the ip address , which causing error on adding the device to the database.
    Resolution:
    The Ipaddress from Vcenter is trimmed before adding to the database.
    (3.7.5, DE428464)
  • Symptom:
    db.passwordEncrypted is a properties key found in the Performance Center webapp properties files twice after upgrade.
    Resolution:
    Instead of appending to the properties files, search and replace with the appropriate value.
    (3.7.5, DE429364)
  • Symptom:
    After proxying to a user and loading user's default dashboard, if the user then brings up the group selection dialog, it will hang when trying to close the dialog.
    Resolution:
    Proxying takes you to the user's default dashboard. The url contains a
    pg
    parameter with no value. That was causing a problem when bringing up the group context selector dialog. We added protective code to handle this case properly.
    (3.7.5, DE429375, 20044304)
  • Symptom:
    Global synchronization for devices could throw error about "Data too long for column 'IPs'" if there were too many devices with same common IP.
    Resolution:
    Removed the function that contained the offending SQL, as it was no longer needed.
    (3.7.5, DE429382)
  • Symptom:
    When a Data Collector goes down and back up during a discovery, the discovery could get into a state of trying to be aborted by the hang detector and not fully get aborted and cleaned up. Resulting in the number of running discoveries to be clogged up by discoveries not running.
    Resolution:
    Added additional info/debug to discovery code around aborting discovery to help track these what is happening.
    (3.7.5, DE430011)
  • Symptom:
    Items are missing from VNA due to persistence failures while processing messages.
    Resolution:
    Updated the VNA persistence error handling to prevent failures from rolling back message processing. Additionally, updated the persistence logic to correctly persist objects that were previously leading to persistence failures.
    (3.7.5, DE430044)
  • Symptom:
    Upgrading Performance Center after previously upgrading to a recent monthly update will appear to be successful, but the dbmigrate log file in the InstallLogs directory will show errors similar to: Sep 04, 2019 11:20:11 AM com.ca.im.installanywhere.util.DbMigrateApp main SEVERE: java.lang.IllegalArgumentException: Expected scheme name at index 0: :mysql\\://localhost\\:3306/netqosportal?characterEncoding\\=UTF-8&useSSL\\=true&verifyServerCertificate\\=false at java.net.URI.create(URI.java:852) at com.ca.im.installanywhere.util.DbMigrateApp.runMigrateMode(DbMigrateApp.java:286) at com.ca.im.installanywhere.util.DbMigrateApp.run(DbMigrateApp.java:164) at com.ca.im.installanywhere.util.DbMigrateApp.main(DbMigrateApp.java:54)  This problem requires two Performance Center upgrades to be exhibited.
    Resolution:
    The Performance Center installer adds escape characters to the database URLs in the various properties files, which were not expected by the DB migration utilities, and were causing errors in parsing such URLs. The DB migration utilities have been updated to properly handle URLs if they have escape characters.
    (3.7.5, DE430601)
  • Symptom:
    The overall performance of Performance Center becomes very slow in a high scale environment when telemetry data is being collected.
    Resolution:
    Update the telemetry database queries to optimize performance.
    (3.7.5, DE430708)
  • Symptom:
    Inventory discovery could hang and then abort when discovering a device that has class D or E IP addresses.
    Resolution:
    Enhanced the device discovery to properly handle IP class D and E IP addresses.
    (3.7.5, DE430883)
  • Symptom:
    When threshold events are forwarded to Spectrum, the quotes around the item are not displayed correctly.
    Resolution:
    Updated threshold event message to use ASCII single quote and double-quotes instead which are displayed correctly.
    (3.7.5, DE431054)
  • Symptom:
    Certain special characters within a custom database password were causing database setup to fail for Performance Center.
    Resolution:
    Updated the installer logic to better handle when special characters are included in the database password.
    (3.7.5, DE431277)
  • Symptom:
    When trying to override the device type for a Meraki device, discovered via SNMP, it would always show Other.
    Resolution:
    Updated the CiscoMerakiManagementMib vendor certification to pass the SysObjectID read from the device to the device type MVEL function. Now it will be able to use the custom DeviceTypes.xml to override the type.
    (3.7.5, DE431612)
  • Symptom:
    Fips is unable to be configured with an encrypted password in the sso.properties configuration file.
    Resolution:
    Update Fips configuration to be able to handle encrypted passwords.
    (3.7.5, DE431709)
3.7.4 Fixes
  • Symptom:
    During install or upgrade, the installation process might fail during a check to see if MySql is running. This is particularly true when using an externalized MySql as we are looking for the running process.
    Resolution:
    The installation process has been improved to check if MySql is running by testing the MySql connection port instead of the mysql process. This allows for the check to validate that MySql is available regardless of whether is it running on the same machine or a different machine.
    (3.7.4, DE413399, 01340717,01337090)
  • Symptom:
    Edit User wizard shows a password error when saving changes to a user using LDAP authentication.
    Resolution:
    Disabled password strength checking when the password fields are disabled.
    (3.7.4, DE415377, 01350459,01349270)
  • Symptom:
    SslConfig could not import key or certificate from a PKCS12 file.
    Resolution:
    With this fix, SslConfig does allow to import key and certificate from a PKCS12 file.
    (3.7.4, DE417649, 01358678)
  • Symptom:
    commons-fileupload-1.2.2.jar is present in the Performance Center Install Directory
    /PerformanceCenter/Performance Center/webapps/Performance Center/WEB-INF/lib/
    . It has been highlighted as a security vulnerability by some sites.
    Resolution:
    Upgrade commons-fileupload to version 1.4.
    (3.7.4, DE419870, 01372590)
  • Symptom:
    Wildfly was configured to use a deprecated version of TLS for encryption.
    Resolution:
    Update Wildfly configuration to allow TLS1.2 for encryption.
    (3.7.4, DE420354)
  • Symptom:
    When choosing to email a dashboard using CSV, it would send more rows than were visible in the viewable pages in the UI. It would use 5000 row max instead of what was configured in the view.
    Resolution:
    Updated the CSV generation to use the view's max row count instead of 5000.
    (3.7.4, DE421872)
  • Symptom:
    When a user with only access to dashboards and Drill into View role right, the menu bar would disappear when drilling into an item due to NullPointerException in
    _topbar.jsp
    .
    Resolution:
    Resolved the issue causing NullPointerException, when drilling into a context page from a dashboard without View Inventory and Search role right.
    (3.7.4, DE422134, 20022262)
  • Symptom:
    Inventory may not update. During upgrade the following log message is seen: 2019-06-27 15:40:17,305 ERROR [org.jboss.as.ejb3.invocation] (Thread-4 (ActiveMQ-client-global-threads-762340938)) WFLYEJB0034: EJB Invocation failed on component InventoryMessageListener for method public void com.ca.em.sdn.gateway.broker.core.inventory.InventoryMessageListener.onMessage(javax.jms.Message): javax.ejb.EJBException: org.hibernate.NonUniqueResultException: query did not return a unique result: 3 at org.jboss.as.ejb3.tx.CMTTxInterceptor.handleExceptionInNoTx(CMTTxInterceptor.java:213) at org.jboss.as.ejb3.tx.CMTTxInterceptor.invokeInNoTx(CMTTxInterceptor.java:265)...Caused by: org.hibernate.NonUniqueResultException: query did not return a unique result: 3 at org.hibernate.internal.AbstractQueryImpl.uniqueElement(AbstractQueryImpl.java:977) at org.hibernate.internal.CriteriaImpl.uniqueResult(CriteriaImpl.java:385) at com.ca.em.sdn.gateway.broker.core.inventory.InventoryPersistenceService.getPersistedCacheKey(InventoryPersistenceService.java:427) at com.ca.em.sdn.gateway.broker.core.inventory.InventoryPersistenceService.removeCacheKey(InventoryPersistenceService.java:344).
    Resolution:
    Updated inventory processing logic to delete multiple instances of cache key entries with the same keystring.
    (3.7.4, DE423273)
  • Symptom:
    User's default group is not accessible via REST.
    Resolution:
    Added defaultGroupId tag to user REST service. Added a new REST endpoint to set default group:
    http://{Performance Center}:8181/Performance Center/center/webservice/users/userId/{User ID}/defaultGroup/{group ID}
    (3.7.4, DE424967, 20028161,20030925)
  • Symptom:
    When the ICMP vendor cert does not back up the device's Reachability and the device is down, the device's status in the context page is shown as Unknown, instead of Down.
    Resolution:
    Improved the polling process to better handle the SNMP reachability so that the device status can be shown correctly.
    (3.7.4, DE425145)
  • Symptom:
    When doing a Data Aggregator synchronization, it may take a long time during PUSH group stage, when there are many site groups using business hours at very large pollable item scale. Checking kill -3 on the data aggregator process during PUSH shows getBusinessHoursID in one of the stacks and repeating after 5 mins and seeing the same stack on the same thread will confirm.
    Resolution:
    Sped up the getBusinessHoursID search greatly by just looking at items with business hour tag on it, and not also checking all items with polllable tag.
    (3.7.4, DE425157)
  • Symptom:
    Inventory discovery hangs when the Data Collector does not return a response to the Data Aggregator. This is due to the data collector throwing an ArrayIndexOutOfBoundsException when it receives invalid SNMP GetNext response, causing it to stop processing instead of returning a response.
    Resolution:
    Fixed the logic to ensure an SNMP Get next response contains the appropriate number of var binds based on the request. If this not the case, the data collector will send a failure response to the Data Aggregator.
    (3.7.4, DE425425)
  • Symptom:
    If a sync push or pull phase for a data source fails due to an exception, there is no END PULL or END PUSH message logged for the sync cycle for that data source. This can make it difficult to understand when each sync phase completes when reviewing the logs.
    Resolution:
    The END PULL and END PUSH messages are logged even if there is a failure during the cycle. The error that caused the failure will be logged shortly after then END PULL or END PUSH log.
    (3.7.4, DE425597)
  • Symptom:
    VNA and CAMM metrics for devices and their components are dropped after using the Stop and then Start Polling actions.
    Resolution:
    We will now correctly recreate the SDN and EMS Polling Configurations when the Start Polling action is sent to a device.
    (3.7.4, DE426041)
  • Symptom:
    Open API does not allow to filter on null values, so it is not possible to create a filter expression find all for example device that have Description property value as null.
    Resolution:
    With this fix, OpenAPI does allow to filter on null value for String property types, filtering on other property type values, such as Double, Integer, is not possible due to OData2 framework limitation. The issue is fixed in Odata4 framework.
    (3.7.4, DE426201)
  • Symptom:
    The Data Collector will not send a poll request if it is determined to by "late". This can happen if the data collector system time moves ahead by at least 90 seconds (if data collector is polling items at 1 minute) or by at least 450 seconds (if data collector is polling items at 5 minutes).
    Resolution:
    The Data Aggregator will generate events indicating that a Data Collector is having this issue, and also generate events on devices for which poll requests are dropped.
    (3.7.4, DE426204)
  • Symptom:
    Performance impact when request 30 days or greater with resolution at poll rate within On-Demand Report, Chart per Item with Multiple Metrics.
    Resolution:
    Addressed the performance impact when request 30 days or greater with resolution at poll rate within On-Demand Report, Chart per Item with Multiple Metrics.
    (3.7.4, DE426440)
  • Symptom:
    When customers updated the password to a value that contained netqos (the original value), multiple replacements would occur causing an invalid password to be stored in the file.
    Resolution:
    Update the replacement logic in the properties file as well as add true password validation during upgrade/install.
    (3.7.4, DE427388)
  • Symptom:
    Description Field for Notifications in MySql is only 255 Char , So when the String to be added is more than 255 char it is throwing an persistence Exception.
    Resolution:
    Increased the Description Field for Notifications to 1024 Char.
    (3.7.4, DE427528, 20030542)
  • Symptom:
    When the Performance Center MySQL instance is installed on a host that is remote to the Performance Center services, firewall rules may close idle DB connections which leads the SSO process to exhaust its connection pool, and eventually hang while processing login requests. This problem should not be present if MySQL is co-located with the Performance Center services.
    Resolution:
    Fixed a defect related to persistent DB connections that would "leak" connections if the connection was closed by an external agent, such as firewalls do with idle connections. Reconfigured DB connection pools used by all Performance Center services to periodically validate idle connections to keep them from being closed by firewalls.
    (3.7.4, DE427880)
  • Symptom:
    When using Spectrum 10.3.2 and VNA Aggregator with Performance Management 3.7, SDWAN tunnel and slapath synced to Performance Center will not consolidate to their Data Aggregator equivalent tunnel and slapath. This is due to the VNA Aggregator changing the SDN ID of the tunnel and slapath to be unique to Spectrum. Performance Center can no longer use it for consolidation. Inventory view for tunnel and slapath will make it appear there are duplicates.
    Resolution:
    Added a new option to Performance Center synchronization code to consolidate new tunnel and slapath on item name only. To enable this option, run in mysql: replace into netqosportal.general values('GlobalSync.SdnConsolidateOnItemName','true');  If you have duplicate tunnel and slapath when using VNA Aggregator, contact support for a SQL script that will clear the item id on Spectrum tunnel and slapath, so they can be re-evaluated for consolidation.
    (3.7.4, DE429278, 20031981)
3.7.3 Fixes
  • Symptom:
    Items would appear in Excluded Items tab for a group after a rule was removed from a group. And could only be added back into the group after removing them from Excluded Items list.
    Resolution:
    Updated SQL to clear the ByRule flag on items removed from a group due to the rule being removed. Items will no longer appear in the Excluded Items list, and can be added back via a new rule.
    (3.7.3, DE403297, 01287437)
  • Symptom:
    Addressed the missing SD-Wan Application Path metric family when customizing a view with context value that is an VNA Site.
    Resolution:
    When building custom view with SD-Wan metric if the context selected is a VNA Site the Application Path metric family is not available in the metric family selector.
    (3.7.3, DE407396)
  • Symptom:
    Cable Modem Maintenance Multi-view was not able to retrieve data for real time status after upgrading to 3.7 release.
    Resolution:
    Addressed query usage when retrieving real time performance data for the Cable Modem Maintenance Multi-view.
    (3.7.3, DE414379, 01341758)
  • Symptom:
    Customers would like option to have the system automatically discover devices contributed through VNA as SNMP manageable.
    Resolution:
    There is now a 'DiscoverVNADeviceAsSnmpManageable' option under the discoverydefaultconfig data aggregator REST endpoint. Setting this option to 'true' will cause the data aggregator to create a discovery profile for the IP Domain with any new VNA IPs.
    (3.7.3, DE415805, 01349698)
  • Symptom:
    Virtual Machines ,Clusters and Hosts were not discovered which were in Folders.
    Resolution:
    Fixed the issue , added logic to discover Inventory in folders.
    (3.7.3, DE416658, 01353535)
  • Symptom:
    When loading the ACI Console at larger scales, it would take the tree a significant time to load the health scores.
    Resolution:
    Updated the ACI Console to load the health scores in a more efficient manner, so the tree renders significantly faster.
    (3.7.3, DE417378, 01356232)
  • Symptom:
    When discovering the Performance Center device via SNMP in the Data Aggregator, you would not be able to view SNMP metrics on the Performance Center item.
    Resolution:
    Updated the Performance Center item to have same default SNMP metric context tabs and ability to add a custom tab of SNMP metrics, when discovered from Data Aggregator.
    (3.7.3, DE417930, 01356609)
  • Symptom:
    On any request OData service responds with "Invalid entity collection or function name 'odata'.".
    Resolution:
    With this fix, if the issue is seen again, the odata server logs full trace back of the exception.
    (3.7.3, DE418095)
  • Symptom:
    A new vendor set
    AWS_IPSEC_TUNNEL
    created for aws tunnel metrics and mapped to SDN Tunnel MF this is turned out to be a default VC for SDN Tunnel MF, which should not be the case.
    Resolution:
    Added <Vendor>Amazon</Vendor> in
    AWS_IPSEC_TUNNEL
    Vendor Cert and updated the version of the SDN Tunnel MF which was missing in earlier delivery.
    (3.7.3, DE418710)
  • Symptom:
    RIB queries that target a Data Aggregator and do not include any GROUP BY, ORDER BY, or LIMIT clause started to fail due to parsing errors on the Data Aggregator after a recent Data Aggregator upgrade.
    Resolution:
    Fixed the Data Aggregator's RIB query parsing logic for this scenario.
    (3.7.3, DE418891, 01367393)
  • Symptom:
    When I select a global collection or enter a string filter value with an ampersand, for example, a global collection named "Demo & Development Systems", I get an error on the alarm console.
    Resolution:
    Modified the code that generates the built in filters, and user defined filters for the alarm console to properly escape values when generating XML requests for alarm data.
    (3.7.3, DE419546, 01368197)
  • Symptom:
    When doing an OData query that had multiple filter groups, it was incorrectly combining the checks to return the wrong items.
    Resolution:
    Updated the OData filter group handling to correctly group the checks for items to find.
    (3.7.3, DE419659, 01366295)
  • Symptom:
    DVS elements were not collected when the they are under multiple network folder structure levels.
    Resolution:
    Fixed the code to support multiple folder structure for DVS.
    (3.7.3, DE419699, 01354247)
  • Symptom:
    Wildfly reload fails on initial Installation.
    Resolution:
    Using Synchronized Version of reload method for wildfly.
    (3.7.3, DE419709)
  • Symptom:
    When running Remote Engineer on Performance Center, it would complain it cannot find
    em.properties
    .
    Resolution:
    Updated the
    re.sh
    script to use the correct path to
    em.properties
    .
    (3.7.3, DE419850)
  • Symptom:
    When upgrading to 3.7, it might fail to run mysql_upgrade program because the password for the account used is not being passed. Check the
    /opt/CA/MySql/mysql_upgrade_results.txt
    file to see if it complained about bad password or other error.
    Resolution:
    Updated Performance Center installer to pass the password to mysql_upgrade program during upgrade.
    (3.7.3, DE420025)
  • Symptom:
    On a scale system Global Search could become too slow since the search criteria is applying a LIKE to every column for each view.
    Resolution:
    With the fix, the search criteria is applying to a given subset of columns for each view.
    (3.7.3, DE420103)
  • Symptom:
    VNA gateway would run in TLSv1.0 or TLSv1.1.
    Resolution:
    Updated VNA gateway to run using TLSv1.2 only.
    (3.7.3, DE420450)
  • Symptom:
    When discovering existing devices and an abort discovery happens before it can complete device discovery and could set the flag that the device was existing, it would delete the existing device.
    Resolution:
    Updated the discovery abort logic to not delete any existing devices.
    (3.7.3, DE420542)
  • Symptom:
    The contrast of link text (a shade of blue) on a selected row in a grid (which has a background color of a pale gray or blue), is insufficient. On certain monitors, the background color is not visible (appears to be white). The shade of blue used for links had a contrast against the background color that is insufficient to meet Accessibility text on background contrast ratio requirements.
    Resolution:
    The background color for selected rows in grids and lists has been darkened and the color used for link text has been adjusted so that it meets accessibility contrast requirements when the link is drawn on the new background color.
    (3.7.3, DE420630)
  • Symptom:
    VNA does not do a refresh of DVS port state before collecting data from vSphere server, Hence reporting same data over for every poll.
    Resolution:
    Added a refresh of DVS Port state before collecting data from the vSphere server.
    (3.7.3, DE421148)
  • Symptom:
    VNA was reporting Virtual Machine and Host Metrics to wrong metric family in data aggregator.
    Resolution:
    Fixed and mapped Virtual Machine and Host metrics to the correct Metric family in data aggregator.
    (3.7.3, DE421149)
  • Symptom:
    The Data Collector can stop receiving inventory and performance data from VNA under the following scenario/\n(1) In Performance Center, go to VNA Gateways admin page, set the Administrative State to Down, and then Up./\n(2) Later on, data collector loses connection to VNA because of network issue or if VNA itself goes down.
    Resolution:
    Fixed logic in Data Collector to ensure it will always restore the connection to VNA when it becomes available.
    (3.7.3, DE421399)
  • Symptom:
    In all supported browsers running on Windows 7 (and likely newer versions of Windows), in all context pages, a horizontal scroll bar appears at the bottom of the page (that scrolls a pixel left or right).
    Resolution:
    In all supported browsers running on Windows 7 (and likely newer versions of Windows), in all context pages, the horizontal scroll bar is always suppressed.
    (3.7.3, DE422022)
  • Symptom:
    During Global Synchronization, if a deadlock or cannot acquire lock issue arises, we would drop temporary tables and remake the connection. This would result when the command is re-run, it couldn't find the temporary table.
    Resolution:
    Updated the retry query handler to not remake the connection on deadlock or can't acquire lock errors. So when command is re-tried, the temporary tables are still there.
    (3.7.3, DE422072)
  • Symptom:
    A VNA device could be mis-reconciled to an SNMP device if the VNA device's primary is found in the SNMP device's secondary IP list.
    Resolution:
    Modified the VNA device reconciliation algorithm to have the Unique ID have a higher precedence than the IP list match.
    (3.7.3, DE422172)
  • Symptom:
    Open API does not show all paths to a group, as a result it's impossible to create a filter to get groups that belong to a path show in Performance Center.
    Resolution:
    With this fix, OpenAPI will show all paths to a group separated by ";".
    (3.7.3, DE422272, 20025914,20025914)
  • Symptom:
    Export the Global Search results to CSV produce a file that is virtually empty.
    Resolution:
    With this fix, , export of the Global Search results would produce a CSV file containing results of the search.
    (3.7.3, DE422647)
  • Symptom:
    After an Event Manager full synchronization is requested, EM spawns it's inventory synchronization. During inventory sync, the group path cache is refreshed. If another EM synchronization is kicked off during this time, it could throw an error that group path cache is being rebuilt.
    Resolution:
    Removed the unneeded check during EM PULL to determine if group path cache was currently being rebuilt. EM synchronization should succeed.
    (3.7.3, DE422795)
  • Symptom:
    Jetty server may expose its details on error pages.
    Resolution:
    with the fix Jetty server is not going to expose details on error pages.
    (3.7.3, DE423311)
  • Symptom:
    Devices with multiple poll rates for same metric family will lose polling after data aggregator restart.
    Resolution:
    Fixed a flaw in the polling configuration recalculation logic that runs during data aggregator startup that caused a loss of polling.
    (3.7.3, DE423637, 20015980)
  • Symptom:
    Performance Center would take a very long time possibly to consolidate SDWAN tunnels and application paths at large scale.
    Resolution:
    Rewrote SQL queries to consolidate SDWAN tunnels and application paths in an acceptable time frame.
    (3.7.3, DE424244)
  • Symptom:
    Addressed issue that Time Tar view is inadvertently locked on to Geo Map context value for SD-Wan Tunnel or Application Path out of the box dashboard.
    Resolution:
    When pinning site context on the GEO-Map for SD-Wan Tunnel or Application Path out of the box dashboard Time Bar view is inadvertently locked on to Geo Map context value.
    (3.7.3, DE424429)
  • Symptom:
    If an existing installation is configured with APM and an upgrade is done to a version that uses OpenJDK, the product will not start.
    Resolution:
    Update APM configuration to be compatible with OpenJDK.
    (3.7.3, DE425273)
  • Symptom:
    The parseSyncTimes.pl script provided with the Performance Center product does not include sync stage timings for data sources that use the HTTPS protocol.
    Resolution:
    The parseSyncTimes.pl script has been modified to report sync stage timings for data sources that use the HTTPS protocol.
    (3.7.3, DE425408)
3.7.2 Fixes
  • Symptom:
    When a device is supported by both CAMM and SNMP the Rediscover and Poll buttons are disabled on the Administration of Monitored Devices.
    Resolution:
    Fixed Rediscover and Poll buttons is disabled on the Administration of Monitored Devices form when device is supported by both CAMM and SNMP on the data source side.
    (3.7.2, DE407753, 01308729)
  • Symptom:
    CABI Scheduled report sent via Performance Center email intermittently arrives with corrupt PDF.
    Resolution:
    Resolved intermittent thread conflict which led to "IllegalStateException: Cipher not initialized." and invalid PDF file.
    (3.7.2, DE411357, 01312291)
  • Symptom:
    Context page takes a long time to load when device is in a lot of groups.
    Resolution:
    Optimized group lookup, and limited context page to ten groups.
    (3.7.2, DE413104)
  • Symptom:
    Depending on network configuration (device, tunnel, interface) the VNA was misidentifying interfaces and relating them incorrectly with devices.
    Resolution:
    Changed logic that relates tunnel source and destination endpoints to interfaces to use unique keys. Added upgrade/migration procedure to fix incorrect keys.
    (3.7.2, DE413335)
  • Symptom:
    After running the
    dr_validate.sh
    script on RedHat 7.x systems, some OS settings are not applied on reboot.
    Resolution:
    Updated the
    dr_validate.sh
    script to mark
    /etc/rc.local
    executable, so Linux will apply them during startup.
    (3.7.2, DE414481, 01345393)
  • Symptom:
    Performance Center Menus disappear when double-quote character " is used in a dashboard's title.
    Resolution:
    A double-quote in a dashboard title is now handled (escaped) and displayed with no negative effect on the Performance Center menu.
    (3.7.2, DE415175, 01345595,01358187)
  • Symptom:
    If the "Drilldown" or "DetailedLogging" parameters are not included in a generated URL in CAPM (such as when a URL from pre-3.6 is used on 3.6+ CAPM), the rendered view shows the "gear" icon and allows editing the view settings.
    Resolution:
    If the "Drilldown" or "DetailedLogging" parameters are not included in URL, CAPM will default to "Drilldown=1" and/or "DetailedLogging=0", and the view cannot be edited.
    (3.7.2, DE415632, 01351680)
  • Symptom:
    Exception on looking up IP address for vm in OC Engine for vsphere plugin.
    Resolution:
    Handled the scenario where IP address was not available for Virtual machine in vsphere plugin.
    (3.7.2, DE416260, 01354780)
  • Symptom:
    Prior to Performance Center 3.7, if there were devices that were contributed by multiple data sources each with a different lifecycle state, then on upgrade to Performance Center 3.7, these device lifecycle state will be inadvertently updated to a common lifecycle state since the upgrade script updates device lifecycle based on consolidated device item id only.
    Resolution:
    Updated the Performance Center 3.7 upgrade script to use device item id along with source id and local id when updating the device lifecycle state.
    (3.7.2, DE417408)
  • Symptom:
    Access Point is shown as Other device instead of manageable device.
    Resolution:
    Updated Access Point as manageable device from other device.
    (3.7.2, DE417857)
  • Symptom:
    Exceptions in the performance and inventory capture causing the inventory and performance not to not populate for the customer.
    Resolution:
    Handled the exceptions in the code, Now the data is populated.
    (3.7.2, DE417918, 01339886)
  • Symptom:
    When Performance Center lays down the Data Aggregator RIB document cache, it is missing all of the metric family RIB documents.
    Resolution:
    Resolved issue with generating metric family RIB documents during build, so they are shipped with Performance Center correctly.
    (3.7.2, DE418508)
  • Symptom:
    Device from spectrum incorrectly reconciles with Device from data aggregator based on primary to secondary ip matching.
    Resolution:
    Modified Primary to secondary matching logic for data sources that provide secondary ip list. Modified reconciliation logic to give higher priority for primary matching and prevent multiple devices from same data sources to get consolidated. If a primary IP matches with more than one devices secondary IP, it should be considered as an invalid match and devices should not be reconciled.
    (3.7.2, DE419176, 01357333,01333613)
3.7.1 Fixes
  • Symptom:
    Filter Report on Monitored Devices screen does not update until after Update Metric Family.
    Resolution:
    Filter Report now updates whenever poll configuration changes.
    (3.7.1, DE262671)
  • Symptom:
    Performance Center's Top Conversations (Pie) view and other NFA related Pie views show a value for
    Other
    that does not match (is smaller than) what NFA reports show.
    Resolution:
    Performance Center's NFA-related Pie views were not correctly summing the values when calculating the "Other" value. This has been corrected and Performance Center's
    Other
    value is now an exact match when using NFA 9.5 and up, and a better (but limited so not perfect) match when using NFA < 9.5.
    (3.7.1, DE392181, 01229503)
  • Symptom:
    When configuring SAML2 for encryption/signing, it would fail to encrypt the SAML2 request correctly if there were more than 1 certificate in the keystore specified in saml.properties.
    Resolution:
    Updated encryption handling in SSO to use the alias specified in saml.properties to grab the correct certificate from the keystore specified.
    (3.7.1, DE403655, 01287800)
  • Symptom:
    Data Aggregator will stop collecting and publishing self-monitoring data in cases where ActiveMQ communication from data aggregator to data collector is not working. It becomes blocked collecting the data collector metrics.
    Resolution:
    The data collector will now publish its self monitoring metrics to PRQ, and data aggregator will collect its metrics as it always has. Self monitoring will no longer be blocked in the case where data aggregator cannot send messages to the data collector.
    (3.7.1, DE404615, 01294587)
  • Symptom:
    When an interface is synchronized by multiple data sources, including Data Aggregator, and user chooses to disable polling, the dialog says it can't be disabled. This is because the code is only checking the first data source it is coming from to see if it's the Data Aggregator.
    Resolution:
    Updated the check, on whether we can disable polling for an interface, to check all data sources the interface comes from, to see if the Data Aggregator is one of them.
    (3.7.1, DE405284, 01270585)
  • Symptom:
    Rollup processing stopped for the NormalizedPortInfo metric family. A data loading thread in the Data Aggregator had been 'stuck' loading into ifstats_rate. Other data loads to ifstats_rate were completing succesfully, but the logic that notified rollup processing was designed only send notifications in the order in which the data load was starting, meaning there were many pending rollup notifications. The only work around to the situation was to restart the Data Aggregator, but all of these notifications would be lost since they were stored in memory.
    Resolution:
    Modified rollup notification logic to send notifications for each data load, regardless of when it was started or if there were other pending data loads.
    (3.7.1, DE405497, 01298195)
  • Symptom:
    When upgrading the Data Collector, it does not prompt to setup the data collector to participate in a Fault Tolerant Data Aggregator environment.
    Resolution:
    When upgrading the Data Collector, it will prompt whether to participate in a Fault Tolerant Data Aggregator environment.
    (3.7.1, DE407376)
  • Symptom:
    During install of the cabi_reports zip file from Performance Center on to the CABI server, if user specified HTTPS scheme, it would save with HTTP scheme.
    Resolution:
    Fixed the installer to correctly save the scheme for Performance Center as HTTPS when specified during install.
    (3.7.1, DE407715)
  • Symptom:
    When a CAMM polling config's item index list contains nulls or is empty, 2 different exceptions could be thrown. One is a IndexOutOfBoundsException in EMSComponentMapper.addComponentMapping. The other is a NullPointerException in MediationSimpleCollector.startPolling.
    Resolution:
    Updated the code to handle if the CAMM polling config item index list is empty or null, so it doesn't throw either of the two exceptions.
    (3.7.1, DE407756, 01312970)
  • Symptom:
    When moving a SNMP device between Data Collectors, the polling config for the CAMM device would have it's associated data collector itemid updated, causing issue with CAMM poll processing. Also, when checking the polling config for a CAMM device, it could show all the SNMP components on the device also.
    Resolution:
    Updated the move device code to not update the CAMM polling config's data collector itemid. Updated the code, that determines what goes into the CAMM polling config, to only include CAMM components.
    (3.7.1, DE407760, 01312985)
  • Symptom:
    When Performance Center sends a trap using nhLiveAlarm format starting with 3.0 GA, it sent the trap varbinds in the wrong OID order compared to the trap definition in the MIB.
    Resolution:
    Fixed the order of the trap varbinds to match the MIB definition for nhLiveAlarm and nhLiveClearAlarm trap definitions.
    (3.7.1, DE408740, 01315252,01315252)
  • Symptom:
    An automated security scan indicated that the version of Jetty used by Performance Center had multiple vulnerabilities that have been fixed in later releases. These automated tests did not allow unauthenticated access to Performance Center.
    Resolution:
    Upgraded the version of Jetty from version 9.4.7.v20170914 to version 9.4.15.v20190215.
    (3.7.1, DE409555, 01321015)
  • Symptom:
    Saved Column location changes are not preserved for "Interfaces List" view when Address column is moved, or when the user's language is changed.
    Resolution:
    Saved Column location changes are now correctly preserved for "Interfaces List" view.
    (3.7.1, DE409671)
  • Symptom:
    Recent modifications to SDWan data source processing injected failures with On-Demand views.
    Resolution:
    On-Demand views need to be updated to to support data source modifications to SDWan data source processing.
    (3.7.1, DE409677)
  • Symptom:
    IM Card View does not support Threshold values more than 10 digits long.
    Resolution:
    IM Card View now supports Threshold values up to 12 digits long.
    (3.7.1, DE410339, 01324894)
  • Symptom:
    Sync intermittently fails with high heap usage.
    Resolution:
    Improved the group path caching mechanism to reduce the memory used by the cache.
    (3.7.1, DE411395, 01331371)
  • Symptom:
    Data Collector may encounter an unexpected exception on critical snmp4j timeout handling thread caused by thread unsafe code and will eventually run out of memory.
    Resolution:
    Made code in Data Collector SNMP on-demand processing thread safe, added better exception handling to ensure that snmp4j timeout handling thread does not fail.
    (3.7.1, DE411528, 01326212)
  • Symptom:
    The pale blue used for some trend charts is not visible.
    Resolution:
    In paired charts such as Interfaces Context Page -> IP Performance -> Interface Utilization/Discard Out and Interface Utilization/Discard In, use the same colors in both charts. Using the paired colors when the In and Out are both shown in the same chart.
    (3.7.1, DE411756)
  • Symptom:
    Dialog hangs when modifying Interface Alias or Speed Override for interfaces that do not report speed.
    Resolution:
    Fixed NullPointerException in dialog.
    (3.7.1, DE411781, 01289974)
  • Symptom:
    The Alarm Console view did not display the Alarm details panel (below the list of alarms) correctly.
    Resolution:
    The Alarm Console view now displays the Alarm details panel correctly.
    (3.7.1, DE412282)
  • Symptom:
    When I click on a site in the SD-Wan GeoMAP in Internet Explorer, I do not see any tunnels or Application paths appear. An error similar to "TypeError - Object doesn't support property or method 'log10'" may appear in the browsers debug console.
    Resolution:
    Internet Explorer does not support a math function log10 that all other browsers support. Modified the code to calculate the base 10 log value of a number on Internet Explorer.
    (3.7.1, DE414275)
  • Symptom:
    Cable Modem Maintenance Multi-view was not able to retrieve data for real time status after upgrading to 3.7 release.
    Resolution:
    Addressed query usage when retrieving real time performance data for the Cable Modem Maintenance Multi-view.
    (3.7.1, DE414379, 01341758)
  • Symptom:
    Cable Modem Maintenance Scorecard Table did not have inventory columns available for selection after upgrading to 3.7 release.
    Resolution:
    Addressed inventory columns not available for selection on the Cable Modem Maintenance Scorecard Table.
    (3.7.1, DE414382, 01341758)
  • Symptom:
    In the Manage Email Settings page, the checkboxes to Enable SSL and to Enable Authentication were not displayed so you could not enable either.
    Resolution:
    In the Manage Email Settings page, the checkboxes to Enable SSL and to Enable Authentication are now displayed and operational.
    (3.7.1, DE414613)
  • Symptom:
    When I give a user the "Send Reports on a Schedule" role right, they do not see the Scheduled Reports item under the Reports menu. They only see it if they also have the On-Demand report template rights.
    Resolution:
    Properly applied the role right checking to show the "Reports" menu if they have the "Send Reports on a Schedule" role right.
    (3.7.1, DE414640, 01344752)
3.7.0 Fixes
  • Symptom:
    The context selector filter for Dynamic Trend View does not properly filter group related items within the view settings dialog.
    Resolution:
    The context selector filter for Dynamic Trend View does not properly filter group related items within the view settings dialog.
    (3.7.0, DE259605)
  • Symptom:
    If a table view was created using the dashboard editor's "Select a context" pinning it to a group context, editing the view would sometimes cause it to revert to the context it was originally created with.
    Resolution:
    As a quick fix, the user can "use defaults", at the tenant level, on the view which should fix the view -or- remove the view, and re-add it, pinning the view outside of the dashboard editor. It turns out that problem is caused because an interaction with the set page size code. We fixed that code, so editing the view will no longer revert/lose the context.
    (3.7.0, DE304922, 00747083)
  • Symptom:
    Vulnerability: CAPerformance Center web UI's JSESSIONID cookie does not have HttpOnly flag set.
    Resolution:
    The JSESSIONID cookie (for both /sso and /Performance Center) now has the HttpOnly flag set.
    (3.7.0, DE362690, 01081094,01248280)
  • Symptom:
    When interface change detection runs on some devices with large amount of interfaces, say 50k, if most of the interfaces have the same ifType value but empty ifDeacr and ifAlias, the change detection will consume huge a mount of memory in the reconciliation process, which then can cause data aggregator shut down.
    Resolution:
    Improved the component reconciliation process to reduce the memory usage for large
    devices.
    (3.7.0, DE362699, 01073787)
  • Symptom:
    Calix Chassis and Blade items can report incorrect status in data aggregator Monitored Devices view because the status is collected by IP Address in the Data Collector.
    Resolution:
    The contact status is now stored in the Data Collector by Device item ID, meaning it is now unique to individual Chassis and Blade items.
    (3.7.0, DE363349, 00993767)
  • Symptom:
    After customer removes a PriorityGroup, it is restored upon upgrade.
    Resolution:
    data aggregator Upgrade will no longer modify PriorityGroups, except for new vendor certifications.
    (3.7.0, DE363592, 01084794,01110569)
  • Symptom:
    The Data Collector would stop collection IPSLA metrics once sysUpTime rolled over on a device.
    Resolution:
    Refactored dynamic index discovery logic to correctly handle sysUpTime rollover and continue collecting metrics.
    (3.7.0, DE364949, 00948977)
  • Symptom:
    The Data Aggregator ActiveMQ web console is shipped and available with unsecured credentials.
    Resolution:
    Provide documentation with instructions on how to disable the Data Collector ActiveMQ web console or to change the users/passwords that can gain access to the console.
    (3.7.0, DE365729)
  • Symptom:
    On-demand report does not honor the context selected in view settings if "Run" is used before "Save".
    Resolution:
    On Demand now applies the context from View settings during "Run", even when the settings have not yet been saved.
    (3.7.0, DE365738, 01083010)
  • Symptom:
    The Data Collector ActiveMQ web console is shipped and available with unsecured credentials.
    Resolution:
    Provide documentation with instructions on how to disable the Data Collector ActiveMQ web console or to change the users/passwords that can gain access to the console.
    (3.7.0, DE366026)
  • Symptom:
    Data Aggregator can return full stack trace in output to an invalid HTTP request.
    Resolution:
    Data Aggregator will not return stack trace in output to an invalid HTTP request.
    (3.7.0, DE366581)
  • Symptom:
    Data Aggregator web services were susceptible to the XML entity expansion vulnerability, for example the billion laughs attack.
    Resolution:
    Turned on the XML parser's secure processing feature.
    (3.7.0, DE366606)
  • Symptom:
    Vendor Cert extension is rejected by web service when IndexTag names are not identical to their AttributeGroup. It is common for an IndexTag name to be the same as its AttributeGroup, but is not a requirement.
    Resolution:
    Fixed validation check in typecatalog web service.
    (3.7.0, DE366901, 01082416)
  • Symptom:
    When doing a SNMP profile discovery using SNMPv2, the device returned noError status with an OID having a syntax of EndOfMibView, and it'd be treated as success. But we'd be unable to poll any OID on the device successfully.
    Resolution:
    SNMP profile discovery now checks the syntax of the OID returned for EndOfMibView and will treat as a failure, and move onto the next SNMP profile, which may be the same community string using SNMPv1.
    (3.7.0, DE367188, 01010475)
  • Symptom:
    Scorecard Trend encounters multiple empty groups when approximate overall average may show "No Data to Display" in the overall column.
    Resolution:
    Addressed potential for "No Data to Display" in the overall column when calculating the approximate overall average with Scorecard Trend with multiple empty groups are found on the system.
    (3.7.0, DE368265)
  • Symptom:
    Exception thrown and context page does not display, when viewing a device that has stopped polling Reachability.
    Resolution:
    Combined the SnmpPollRate and PollRate enums in CAPerformance Center, so that the context page can handle a stopped Reachability. Updated context page UI to display a notification when all views on the page are suppressed.
    (3.7.0, DE368317, 01069507)
  • Symptom:
    The 'Start Poll' button in Data Aggregator Monitored Devices view for a Calix Blade does not fully start polling.
    Resolution:
    The 'Start Poll' button in Data Aggregator Monitored Devices view for a Calix Blade will fully start polling.
    (3.7.0, DE369062, 01110985)
  • Symptom:
    When I administer a tenant, proxy a user, and then create a dashboard, I will get a pop up message stating "All views on this page are suppressed; redirecting to your Home page.". I can not see the menu until I stop proxying and proxy the user again.
    Resolution:
    Modified the code responsible for refreshing menus to properly detect when a user is being proxied. (3.7.0, DE369374)
  • Symptom:
    When using DiscoveryProfile's Add attribute to add IP/hostname to the discovery profile multiple times with the same values, the IP/hostname could not be added.
    Resolution:
    Improved the discovery profile incremental Add function so that it will always add the IP/hostname to the profile.
    (3.7.0, DE370047, 01112133)
  • Symptom:
    When re-configuring the Data Aggregator/Data Collector info in the DB using the
    update_da_dc_database_references.sh
    script, it would not find the DA/data collector device items if they were also SNMP managed devices.
    Resolution:
    Modified the
    update_da_dc_database_references.sh
    script to handle looking for the data aggregator/data collector device items even when SNMP managed. Check the script and documentation for new variables to update before running script.
    (3.7.0, DE370063, 01074972)
  • Symptom:
    CAPerformance Center REST API for getting list of groups sometimes returns incomplete results, when other Group Edit activities like running Group Rules cause the group cache to be rebuilt.
    Resolution:
    CAPerformance Center REST API for getting list of groups has been modified to retry up to 5 times when it encounters an invalidated group cache, so it will now return a complete list.
    (3.7.0, DE370724, 00967870,00967870)
  • Symptom:
    Scheduled CSV output for trend charts does not support full data in report when with On-Demand Report and Dynamic Trend View.
    Resolution:
    Addressed the limitation with scheduled CSV reports not containing all data when generated with trend charts for On-Demand Reports and Dynamic Trend View.
    (3.7.0, DE371031, 01104216)
  • Symptom:
    Upgrades do not preserve the saml keystore file specified in the saml.properties entry saml.sp.certificate.location if the keystore file is not located in the same directory as saml.properties.
    Resolution:
    Modified the installer to read the location of the keystore file from saml.properties, back up, and then restore the specified file as part of the upgrade install.
    (3.7.0, DE371073, 01121935)
  • Symptom:
    Under certain conditions, the data collector can exhibit decreased performance, increased memory growth and unhandled exceptions.
    Resolution:
    The data collector has been improved; it will be more stable and provide better logging for exceptions.
    (3.7.0, DE371213, 01110803)
  • Symptom:
    Dashboards that depend on finding a list of Tunnel or SLAPath components are extremely slow at scale.
    Resolution:
    Re-worked these queries, that gather the Tunnels or SLAPath, to use the available indexes to reduce these queries to under 10 secs at high scale, and sub-second in low scale.
    (3.7.0, DE371465, 01117584)
  • Symptom:
    Pie charts are not displayed when receiving non-numeric data for a metric from a data source. Log indicate a NumberFormatException is occurring.
    Resolution:
    Updated the pie chart creation to handle any metric values it cannot convert as 0.0, so the pie chart will display.
    (3.7.0, DE372015, 00995663,01147886)
  • Symptom:
    1. On CA Performance Center systems that have events on items with unicode characters, the upgrade to version 3.6 fails. The install logs will show an error similar to the following:/Error executing command: CALL item_grouppath_tables() on line 223 in script /opt/CA/PerformanceCenter/EM/SQL/upgrade/3.6.0.0/update_grouppath_tables.sql Incorrect string value: '\\xDE\\x8F' for column 'ItemName' at row 1./2. On a CA Performance Center system that is already running version 3.6, events may fail to be added to the Event Manager database if some of the events are on items with unicode characters. The EMService.log file will show errors similar to the following:/Incorrect string value: '\\xDE\\x8F' for column 'ItemName'.
    Resolution:
    The default character set of the new event_item_info and item_changes_work tables had been set to 'latin1'. This led to errors if item names included unicode characters. The resolution is to set the default character set to UTF8 on the tables, and to change the character set to UTF8 on the problematic columns.
    (3.7.0, DE372312, 01127164)
  • Symptom:
    When Performance Center is running in HTTPS mode, links to Application Delivery Analysis required the user to login to ADA. This is because the page didn't have access to SSO token to pass to ADA.
    Resolution:
    Updated the SSO token cookie flag to allow access to it by client in order to pass to ADA and not require login.
    (3.7.0, DE373999, 01130796,01130789)
  • Symptom:
    When attempting to save the "Threshold Monitoring" tab on the data aggregator context page within the Page Builder there is an error "You must supply a tab title".
    Resolution:
    Addressed the error "You must supply a tab title" when attempting to save the "Threshold Monitoring" tab on the data aggregator context page within the Page Builder.
    (3.7.0, DE374716)
  • Symptom:
    Customer vendor certifications and metric families caused a RIB error on dashboards.
    Resolution:
    Updated the dependencies in the project to pull in the correct version of vendor certifications and metric families so that customizations are honored as expected.
    (3.7.0, DE375196, 01140488)
  • Symptom:
    When a threshold profile contains multiple rules for the same metric family, a SQL exception is thrown and the evaluations are not made.
    Resolution:
    Fixed the dynamic SQL query generated, so it processes the event evaluations without error.
    (3.7.0, DE375239)
  • Symptom:
    On a SNMP and CAMM reconciled device, if a metric family is supported by both SNMP and CAMM, the CAMM components could be marked as Not Present when the metric family's SNMP change detection runs.
    Resolution:
    Improved the reconciliation process to better handle components from different protocols when SNMP change detection runs.
    (3.7.0, DE375970)
  • Symptom:
    The Data Aggregator data directory is always set to the default (/opt/IMDataaggregator/data) in non-HA fresh install.
    Resolution:
    Set the Data Aggregator data directory based on user specified install directory.
    (3.7.0, DE376293, 01145499,01151619)
  • Symptom:
    When viewed in the context of a group with more subgroups than the configured limit, CA Performance Center IM Group Scorecard Table and IM Group Scorecard Trend views may show different and/or incorrect values in subgroup data. Top-N subgroups for the same parent group and time range may be different each time the views are rendered.
    Resolution:
    The Data Aggregator did not support RIB queries that group by group ID. This required CA Performance Center to make individual queries for each subgroup. The number of subgroups for which CA Performance Center would make such queries was governed by a configurable limit. This problem could be seen if the number of subgroups exceeded the limit. The Data Aggregator now supports RIB queries that group by group ID and CA Performance Center uses such queries in the IM Group Scorecard Table and IM Group Scorecard Trend views to ensure the correct Top-N subgroups and corresponding metric values are presented every time.
    (3.7.0, DE376472, 01146769,00964694)
  • Symptom:
    When editing a On-Demand Report under the Reports menu the view settings are not always saved.
    Resolution:
    Addressed the problem with view settings not always being saved when editing On-Demand Report under the Reports menu.
    (3.7.0, DE376970)
  • Symptom:
    First page of PDF is blank and some views are clipped on a dashboard containing two columns of multiple Calendar Heat Chart views each, and some of the views have "No Data".
    Resolution:
    PDF output of 2-column Calendar Heat Chart views no longer have blank first page, and the "No Data" views align better with the views having data, yielding better page breaks.
    (3.7.0, DE376986, 01136276)
  • Symptom:
    A device Context Page is missing all the tabs and views associated with the device type. The context page is basically blank except for the context header (device name). An exception was emitted in the Performance Center Service log.
    Resolution:
    Validation code was added to verify that the correct page was retrieved and was not null. This eliminated the exception in the Performance Center Service log and all the tabs and views are now displayed correctly.
    (3.7.0, DE377273, 01147601,01216731)
  • Symptom:
    When trying to clear an entry in the rollup cache in the Data Aggregator, it would display a message about "invalid cleanup statement".
    Resolution:
    Fixed the parsing code of the resetRollup config operation, so it will clear the correct cache entry. Added better error messaging to indicate if we failed to parse or the QName entry was malformed.
    (3.7.0, DE377741)
  • Symptom:
    A Performance Center initiated upgrade of a data collector from 3.5 to 3.6 in a Fault Tolerant configuration will fail with an exception in the data aggregator log.
    Resolution:
    A Performance Center initiated upgrade of a data collector in a fault tolerant configuration will now succeed.
    (3.7.0, DE377936)
  • Symptom:
    When upgrading the Data Collector, that is installed into the non-default install location, install would complain about being unable to find the apache-activemq-0/conf/activemq.xml file.
    Resolution:
    During data collector upgrades, install will use the existing install directory when determining the installed ActiveMQ version, so it knows whether ActiveMQ needs to be upgraded.
    (3.7.0, DE378481, 01155766,01156367,01239285)
  • Symptom:
    When trying to load views on SDN-WAN dashboards, and no data is seen, a SQL error is seen in logs like: Unknown column 'im.ParentItemSubType' in 'where clause'.
    Resolution:
    Fixed SQL call being made to use correct column from item_members table.
    (3.7.0, DE378981)
  • Symptom:
    An automated application scan indicated a Blind SQL Injection vulnerability related to a parameter in the Performance Center URL. This vulnerability did not allow unauthenticated access to Performance Center or access to information the user would not normally see. Normally this parameter contains an id for the currently selected menu.
    Resolution:
    Added validation related to the parameter in the Performance Center URL.
    (3.7.0, DE379741)
  • Symptom:
    An automated application scan indicated a Blind SQL Injection vulnerability related to a parameter in the Performance Center URL. This vulnerability did not allow unauthenticated access to Performance Center or access to information the user would not normally see. Normally this parameter contains an id for the Data Aggregator.
    Resolution:
    Added validation related to the parameter in the Performance Center URL.
    (3.7.0, DE379742)
  • Symptom:
    An automated application scan indicated a Cross-Site Scripting vulnerability related to the page parameter in the Performance Center URL. This vulnerability did not allow unauthenticated access to Performance Center or access to information the user would not normally see. Normally this parameter contains a page id or a context id for a context page.
    Resolution:
    Added validation related to the page parameter in the Performance Center URL.
    (3.7.0, DE379744)
  • Symptom:
    Reports that are scheduled to send email on the first day of the month, in a month after a "short" (<31 days) month, for a UTC+n timezone, may be delivered a day early, and then repetitively, in the following month.
    Resolution:
    Calculation of offsets for UTC+n timezones (for example Tokyo), have been corrected. Emails scheduled for the first day of the month in UTC+n time zones now correctly send on the first day of the month as expected.
    (3.7.0, DE380620, 01154925,01154762)
  • Symptom:
    CAPerformance Center Dashboard view that has business hours applied shows no data when "Last 30 days" time range is selected.
    Resolution:
    CAPerformance Center was incorrectly requesting daily data from the Data Aggregator in this case. It has been corrected to request hourly data and now shows data that is filtered to the specified business hours.
    (3.7.0, DE380775, 01158861)
  • Symptom:
    Data Aggregator would experience memory growth related to a backup of messages in shared Distributed Item Repository thread pool.
    Resolution:
    (1) Segregate Distributed Item Repository work load by data collector, and limit the number of messages in each queue./(2) Stop excessively opening/closing data aggregator side distributed item repository message producers to avoid unnecessary work in the broker, and can lead to thread deadlock in the Data Aggregator.
    (3.7.0, DE380783, 01164122,01172789,01150084)
  • Symptom:
    Monthly Scorecards for Previous Month return data aligned to the middle of the month and assume that each month is 30 days long. This is a data integrity issue.
    Resolution:
    The Monthly Scorecard for Previous Month, Current Month and a Custom Time Range of a month will now return data aligned to the end of the previous month and each Month output column will aggregate results according to the correct number of days within the Month displayed. This fix ONLY fixes the data integrity issue for Monthly reporting.
    (3.7.0, DE381296, 01166376,01175677,01261459)
  • Symptom:
    With 200+ tenants, an exception (as given in the description section) would be thrown while trying to store tenant relationships (with out of box collections) as a scalar attribute in vertica db.
    Resolution:
     The original set of multikeys is now stored in the vertica db as 4 different list attributes - ( RelationshipQName, LeftItemId, RightItemId, TenantId ).
    (3.7.0, DE382445)
  • Symptom:
    Performance Center synchronization may report SQL deadlock issues, due to use of InnoDB locking.
    Resolution:
    Changed the InnoDB transaction-isolation to READ-COMMITTED, to greatly reduce the chance of deadlock.
    (3.7.0, DE382458, 01175801,01172789,01195867)
  • Symptom:
    An Open API query to get SDN tunnel metric family data with a group filter may return an empty result set if parent SDN device is not added to the group and the group include children flag is not set. For example odata/api/sdntunnelmfs?&$select=ID,im_Latency&$filter=(groups/Name eq 'Tunnel Group').
    Resolution:
    With this fix, the OpenAPI will return data for directly added and children sdn tunnels of a group.
    (3.7.0, DE383159)
  • Symptom:
    User settings in OnDemand Views, such as "Max-Per-Page" and saved column selections, are cleared when the view is edited and saved from the dashboard [age editor.
    Resolution:
    User settings in OnDemand Views, for "Max-Per-Page" and saved column selections, are now preserved when the view is edited and saved from the dashboard [age editor.
    (3.7.0, DE383163, 01174568,01190643)
  • Symptom:
    When dashboard is scheduled as "Now" with a custom time range time is not properly honored in the generated report.
    Resolution:
    When dashboard is scheduled as "Now" the generated report properly honors the with a custom time range time.
    (3.7.0, DE383811, 01179804)
  • Symptom:
    Reports for "Switches With Highest Memory Utilization (Table)" shows blank on 3.6.
    Resolution:
    The problem is general switch views accidentally used "ACI specific rib query", therefore, it will only return ACI switches. After fix, Reports for "Switches With Highest Memory Utilization (Table)" works fine now.
    (3.7.0, DE384381, 01166064)
  • Symptom:
    When I save column settings for an inventory list as an administrator, users who have the "view item name alias only" role right see the real name for the item, not the alias.
    Resolution:
    Modified the code to enforce the "view item name alias only" role right for inventory lists where an administrator has saved column settings.
    (3.7.0, DE384690)
  • Symptom:
    If saml.properties is configured in PerformanceCenter to use metadata file outside of /opt/CA/PerformanceCenter/sso/webapps/sso/configuration, during upgrade, the metadata file will not be restored. Work was already done to backup/restore the keystore file.
    Resolution:
    Updated installer to backup the metadata file, and restore the file during upgrade.
    (3.7.0, DE384880)
  • Symptom:
    Group relationships might be created in Data Aggregator, during synchronization, where the member's item_id no longer exists in the DA, causing extra data aggregator memory and DB storage to be used. This could happen if items were deleted from data aggregator DB directly instead of going through REST or UI.
    Resolution:
    Data Aggregator will confirm the member's item_id exists in the data aggregator before allowing the relationship to be created.
    (3.7.0, DE384887, 01185918)
  • Symptom:
    Devices with item names longer than 100 characters failed to sync to CAPerformance Center.
    Resolution:
    Updated sync to trim down device item name sizes longer than the size that can be handled in CAPerformance Center and increased acceptable device item name sizes to 255 characters.
    (3.7.0, DE385035, 01255096)
  • Symptom:
    Rollup processing can be delayed in large environments as there is a limitation of only consuming 400 messages per minute from a JMS rollup queue.
    Resolution:
    Changed the ActiveMQ maxBrowsePageSize setting from default of 400 to 10000 to better keep up with Rollup message processing.
    (3.7.0, DE385203, 01191652)
  • Symptom:
    The CA Business Intelligence Reports list is not the height of the screen, like the Inventory, On-Demand Report Templates, and other lists.
    Resolution:
    Modified the code so that the CA Business Intelligence Reports list is a "full height" list, like other similar lists.
    (3.7.0, DE385418, 01193292)
  • Symptom:
    When updating the Event Manager hostname/IP on the disaster recovery system, the synchronization process would not send the new hostname/IP to Event Manager. When Event Manager re-register itself with the Data Aggregator, it would use the wrong hostname/IP.
    Resolution:
    Updated the script that modifies the Event Manager hostname/IP to also update the UpdatedOn timestamp field with current date/time, so it is synchronized to Event Manager.
    (3.7.0, DE386125, 01189790)
  • Symptom:
    When Performance Center requests specific time frames for group trend scorecard (Monthly), the Data Aggregator currently attempts to widen the range in the past to align with the resolution.
    Resolution:
    Data Aggregator now respects the user specified start time for a monthly group trend scorecard.
    (3.7.0, DE386139)
  • Symptom:
    Cisco Nexus 5596UP is discovered as Cisco Nexus 5020T.
    Resolution:
    Updated ModelMap.
    (3.7.0, DE386569, 01163091)
  • Symptom:
    Data Collector processing of CAMM inventory and polling XML files could be killed, if it took longer than 5 minutes to process both files.
    Resolution:
    Increased the default processing time to 15 minutes and made the setting configurable. Create file com.ca.im.dm.mediation.platform.impl.MediationSimpleCollector.cfg file and add: processingTimeout=VALUEINMILLISECONDS.
    (3.7.0, DE386759, 01197065)
  • Symptom:
    When using detailed projections with group scorecard trend there are scenarios that double icons showing first projected data column.
    Resolution:
    Addressed the possible scenario there are double icons showing first projected data column using detailed projections with group scorecard trend view.
    (3.7.0, DE387155)
  • Symptom:
    Sort direction arrow in Group Scorecard Table View shows wrong direction when view is first rendered.
    Resolution:
    Sort direction arrow in Group Scorecard Table View is initialized to show the correct direction, as defined in the view's settings.
    (3.7.0, DE387563, 01203493)
  • Symptom:
    When configuring an on-demand report, clicking the add/remove items button can take a long time depending on group structure and how many different itemtype/subtypes have items.
    Resolution:
    Added a workaround to just check if any item has the itemtype/subtype, instead of just what the user has access to. Run this to use the override: REPLACE INTO netqosportal.general VALUES ('OnDemand.ItemTypeUseAllGroups','true');.
    (3.7.0, DE387585, 01193773)
  • Symptom:
    The ADA view named "Top Performance Map by Server (Bar Chart)" shows "No Data to Display" for all cells in the column that normally shows percent bar for the selected metric type.
    Resolution:
    The ADA view named "Top Performance Map by Server (Bar Chart)" now shows percent bar for the selected metric type.
    (3.7.0, DE387854, 01199275)
  • Symptom:
    QueryBuilder does not populate the discoveryprofile/UseListOfSnmpProfiles attribute in the HTML Table format when used in an expand query.
    Resolution:
    With this fix, QueryBuilder does populate the discoveryprofile/UseListOfSnmpProfiles attribute in the HTML Table format when used in an expand query.
    (3.7.0, DE388007)
  • Symptom:
    When doing a percentile report, the order of the data may be incorrect due to NULL values in the metric used for sorting.
    Resolution:
    Modified the SQL query to conform to how vertica handles sorting data with NULLs. It will now order the data so NULLs are always last in the sort order.
    (3.7.0, DE388766, 01212746)
  • Symptom:
    Performance Center synchronization process could hit a deadlock or lock timeout trying to write to InnoDB tables.
    Resolution:
    Updated database query execution to check for the deadlock and lock timeout occurring and retry the command up to 2 times more.
    (3.7.0, DE389072, 01213752)
  • Symptom:
    SD-WAN GeoMap would throw an error getting performance data, because it didn't find any tunnels or slapaths to report on for the Site group(s) selected.
    Resolution:
    Fixed SQL query to search for tunnels and slapaths for the Site group(s) selected.
    (3.7.0, DE389760, 01212034,01217861,01220040)
  • Symptom:
    Performance Center group selection sometimes resets to user's default group after group id is passed on a URL like this: "http://yourPerformance CenterHost:8181/Performance Center/desktop/page?GroupID=1500" (where 1500 is id of group being selected).
    Resolution:
    Group selection made on URL can now be persisted by using the new "SetLastGroupID" parameter in the URL, like this: "http://yourPerformance CenterHost:8181/Performance Center/desktop/page?GroupID=1500&SetLastGroupID=true" (where 1500 is id of group being selected).
    (3.7.0, DE389939, 01193711)
  • Symptom:
    Depending upon the theme selected the Scorecard trends show broken icons with invalid data is returned.
    Resolution:
    Addressed the broken icon shown on Scorecard trends with invalid data returned with custom theme selected.
    (3.7.0, DE389970)
  • Symptom:
    When running the silent installer for fault tolerant Data Aggregator, after using the installer to generate the response file, it may fail to start the consul service due to double quotes in the config.json file. The installer generated a bad response file.
    Resolution:
    Updated the data aggregator installer to generate a valid response file for fault tolerant systems. If you have a response file already, please run the installer to regenerate the response file.
    (3.7.0, DE389996, 01205561)
  • Symptom:
    Event manager inventory requests to device manager may fail with errors in the device manager logs that point to multiple event manager instances accessing the same device manager. This can happen if Performance Center is configured for Disaster Recovery and the backup server starts while the primary is running, however there is no way to diagnose the source of the inventory service calls from the device manager logs.
    Resolution:
    All jetty services (console, device manager, and event manager) will now log web requests, including source address, and retain request logs for three days.
    (3.7.0, DE390322)
  • Symptom:
    When selecting Select Context link in dashboard/context page editor, or Add/Remove Items button in on-demand reports, it can take minutes for the dialog to come up with the list of item type/subtypes the user has access to. It may even time out.
    Resolution:
    Re-wrote the SQL queries that determines what items are in the group hierarchy the user has access to. The new queries will use existing indexes better, and will result in the dialog coming up faster. Inventory views may load faster as they use common SQL queries.
    (3.7.0, DE390341, 01193773)
  • Symptom:
    Business Hour filter may not be reflected in view subtitle as to processing of data source resolution property Scorecard Trend view.
    Resolution:
    Addressed data source resolution property in Scorecard Trend view may not be reflected Business Hour filter in subtitle.
    (3.7.0, DE390524)
  • Symptom:
    Multi-Trend removed the parent device context link showing link unavailable on component level chart for non-synchronized items in 3.6 release.
    Resolution:
    Added back the parent device context link with Multi-Trend component level chart for non-synchronized items based on customer feedback.
    (3.7.0, DE390537, 01215036)
  • Symptom:
    When clicking on one of the severity bubbles in Card view to show the items, it may show no item name or wrong item name. Due to the data aggregator data source item id being equal to the performance center item id for a different item.
    Resolution:
    Updated the item id translation code, in Performance Center, to check if the field coming back in the data from the data aggregator is the data aggregator item id or a performance center item id, and do the correct lookup to show the item to display.
    (3.7.0, DE390748, 01212597)
  • Symptom:
    The Out Of Box Event Rule 'Polling Safety Value' can be disabled by customer, but if data aggregator restart occurs, the event rule will be enabled.
    Resolution:
    1) Changed the Event Rule data model to make the 'Enabled' field option, and set to true by default\u003Cdiv2) Changed the apache-karaf-2.4.3/xml/eventrules/PollingSafetyValveRule.xml to include Enabled, so it will not be overwritten by functionality that creates/updates EventRule item at data aggregator startup.\u003C/div.
    (3.7.0, DE391365, 01219035)
  • Symptom:
    Group Scorecard Trend for Previous Month not showing last day of month (e.g. Sept 29 instead of Sept 30) in column header when the user's timezone is west of (has negative offset from) the DR system time zone.
    Resolution:
    Group Scorecard Trend for Previous Month shows correct last day of month in column headers regardless of the user's time zone.
    (3.7.0, DE391422, 01166376)
  • Symptom:
    OpenAPI does ignore group's "do not include children" flag for queries on /components with a filter of Group/Name = group. It shows all components of parent devices not just directly added components.
    Resolution:
    With this fix, OpenAPI will honor group's "do not include children" flag for queries on  /components with a filter of Group/Name = group, and show only devices directly added to the group.
    (3.7.0, DE392172, 01225937)
  • Symptom:
    After using data collector UI upgrade feature, the data collector can remain in Upgrading state even after data collector upgrade has completed, has reconnected to DA, and appears functional.
    Resolution:
    Fixed the CollectorState handling logic to ensure data collector transitions to Collecting Data after upgrade is complete.
    (3.7.0, DE392398)
  • Symptom:
    After upgrading Performance Center to 3.6.2 or later from 3.6.0 GA, Performance Center failed to apply the 3.6.1 SQL DB upgrades.
    Resolution:
    Moved the 3.6.1 upgrade SQL DB calls from 3.6.0.0 SQL files to 3.6.1.0 SQL files, so they are run coming from 3.6.0.0 to 3.6.2 or later.
    (3.7.0, DE392586)
  • Symptom:
    Group Scorecard Trend view with Business Hours site returns incorrect intervals with 'No Data to Display'.
    Resolution:
    Group Scorecard Trend view no longer shows incorrect intervals with 'No Data to Display'.
    (3.7.0, DE392851)
  • Symptom:
    Performance Center Single Sign On audit logging used to create 1 CSV line per login request, but is no longer.
    Resolution:
    Updated SSO audit logging to store 1 CSV row per login request.
    (3.7.0, DE393523)
  • Symptom:
    Users cannot use the
    update_alias_name.sh
    script when Performance Center is running HTTPS.
    Resolution:
    Updated the
    update_alias_name.sh
    script to support Performance Center running in HTTPS mode.
    (3.7.0, DE393588)
  • Symptom:
    When I edit a dashboard that contains an OpenAPI App View, it sometimes gets locked on a particular item. There appears to be no way to change it back to being dynamic.
    Resolution:
    Added the ability to set or clear the context for an OpenAPI App, similar to other Performance Center views. This allows OpenAPI apps to be set to a specific group, device, or other item context. It also allows contexts to be cleared, similar to other Performance Center views.
    (3.7.0, DE393757, 01239213)
  • Symptom:
    When installing Performance Center on SLES 12 SP2, InstallAnywhere is unable to determine free disk space available correctly. Using -DCHECK_DISK_SPACE=OFF only helps with the main Performance Center installation, but the modules it calls to install MySQL, DM, EM, Performance Center, and SSO still do free disk space checks and fail to install the product.
    Resolution:
    Updated the modules and Performance Center installer to propagate -DCHECK_DISK_SPACE=OFF so they don't check for free disk space and install the product fully without errors.
    (3.7.0, DE394348, 01236497)
  • Symptom:
    Customers with a larger number of Tenants experience slow page rendering.
    Resolution:
    Added an index to speed up the query used in rendering pages.
    (3.7.0, DE395353)
  • Symptom:
    Developers on Windows are unable to use the
    npcshell.bat
    script.
    Resolution:
    Change git attributes to convert line endings for .bat files to use CRLF.
    (3.7.0, DE395397)
  • Symptom:
    After installing/upgrading a fault tolerant data aggregator, and then the second DA, the first data aggregator might report an error about relation v_group does not exist. This is due to a DB change being applied during the second data aggregator install that should not have been.
    Resolution:
    Updated data aggregator installer to only run the data aggregator view creation/recreation SQL, when it's the first data aggregator in a fault tolerant environment being installed/upgraded.
    (3.7.0, DE395611)
  • Symptom:
    Performance Center EM notification script sometimes throws concurrent modification exception when scripts and email are both executed for same event.
    Resolution:
    Updated the common property binder map used by both email and scripts so that it can be accessed concurrently.
    (3.7.0, DE396497)
  • Symptom:
    Custom Trend view when specify the Stacked Chart option was not rendering first metric on chart.
    Resolution:
    Addressed Stacked Chart option was not rendering first metric on chart on Custom Trend view.
    (3.7.0, DE396693)
  • Symptom:
    The vendor certs that have VCSupportExpression and keyed attributes could be marked as not supported if a device's SNMP agent does not return the SNMP responses properly.
    Resolution:
    Improved the SNMP reading process to better handle the keyed attributes on the vendor certs that have VCSupportExpression.
    (3.7.0, DE397037, 01233458)
  • Symptom:
    "Top Interface Utilization - MultiTrend (Trend)" view does not honor Direction Settings = "Out" / "In" in view settings - always sorts by Utilization.Avg.
    Resolution:
    "Top Interface Utilization - MultiTrend (Trend)" view now honors the user's "Direction Settings" choice and sorts by UtilizationOut.Avg or UtilizationIn.Avg, as appropriate.
    (3.7.0, DE397196, 01258722)
  • Symptom:
    When fields are modified and saved in User Settings, the
    Default Group
    resets to My Assigned Group and
    Time Display Format
    resets to 12 Hour Format.
    Resolution:
    Preserve settings for
    Default Group
    and
    Time Display Format
    when other fields are modified and User Settings are saved.
    (3.7.0, DE398245, 01263751)
  • Symptom:
    On RedHat 7 the Performance Center MySQL service is not enabled so it does not automatically start after restarting the machine.
    Resolution:
    The Performance Center installer has been updated to enable the MySQL service when run on RH 7.
    (3.7.0, DE398639)
  • Symptom:
    Top Interface Utilization "In&Out" sorts by "UtilizationIn ASC" - should sort by "Utilization DESC".
    Resolution:
    Top Interface Utilization views, and similar views for Errors and Discards, now sort by the "In" or "Out" metric when selected, or by the matching total metric, e.g. "Utilization.Avg" when "In&Out" is selected.
    (3.7.0, DE400601, 01263271,01290627)
  • Symptom:
    The Life Cycle State for devices and interfaces may have an incorrect case in the database.
    Resolution:
    Corrected issues with DB schema management that maintains tables related to synchronization. Fixed up incorrect data at upgrade time.
    (3.7.0, DE400650)
  • Symptom:
    The device rest service function to modify the life cycle state of a device allows you to change the life cycle of the Data Aggregator and Performance Center item.
    Resolution:
    These are special items that do not support life cycle management. The web service has been modified to return an error when an attempt to alter the life cycle of these items is performed.
    (3.7.0, DE400830)
  • Symptom:
    After I setup SSL on the Data Aggregator and import the certificate on the Performance Center system, I do not see data in the views. I see a warning in the Data Aggregator karaf log stating: javax.net.ssl.SSLException: Received fatal alert: certificate_unknown.
    Resolution:
    After importing the Data Aggregator certificate, the device manager process needs to be restarted. Modified SSlConfig to restart Device manager and the Performance Center console when importing the Data Aggregator certificate.
    (3.7.0, DE401059)
  • Symptom:
    I manually configured Performance Center to use HTTPS. When I attempt to log in, I get the message "The application associated with this instance of Single Sign-On does not appear to be running". The SSL Health check in SslConfig does not report any issues.
    Resolution:
    Log on to Performance Center is configured by a file named CAPerformanceCenter.xml located in /opt/CA/PerformanceCenter/sso/webapps/sso/configuration. There is also a file that configures log on to OData Query named CADataAggregator.xml. Since Single Sign On can be extended by adding other configuration files, Each file contains a product code that identifies it as either being Performance Center or Data Aggregator. If a copy of one of these files is made with the extension .xml, for example CAPerformanceCenter-backup.xml, it can be picked up by Single Sign On. The SSL Health Check will now search for, and report on duplicates of CAPerformanceCenter.xml and CADataAggregator.xml. It recommends that these duplicates be removed, or moved to a backup directory.
    (3.7.0, DE401397)
  • Symptom:
    IP Domain fails to create for unspecified reason.
    Resolution:
    Update the database connection used for temporary tables for IP domain creation.
    (3.7.0, DE401609)
  • Symptom:
    When any device gets de-consolidated (maybe due to an IP address change), we see that the old device to interface relationships do not get deleted.
    Resolution:
    Update CAPerformance Center sync code to delete old device - interface relationships.
    (3.7.0, DE401659)
  • Symptom:
    Performance Center system status page for Data Collector was failing to load in 20 secs due to busy Data Aggregator, causing it to show Failed.
    Resolution:
    Increased the time Performance Center waits for the data aggregator to return the Data Collector status information.
    (3.7.0, DE401890, 01277900)
  • Symptom:
    Update actions in CAPerformance Center fail for an unspecified reason intermittently.
    Resolution:
    Update database connection type for these transactions.
    (3.7.0, DE402053)
  • Symptom:
    Self Monitoring metrics stop by collected when Data Aggregator is busy, as seen by a ConcurrentModificationException in data aggregator
    karaf.log
    .
    Resolution:
    Refactoring self monitor poller to correctly handle concurrent invocations.
    (3.7.0, DE402065, 01281322)
  • Symptom:
    The Data Aggregator
    file-monitor.sh
    script throws an error when the shared directory is not under the default location.
    Resolution:
    Updated the
    file-monitor.sh
    script to first determine where the shared data directory is located, and use that value to monitor the DTO files in performance-spool.
    (3.7.0, DE402286)
  • Symptom:
    Life cycle state of some Spectrum devices do not get updated when spectrum synchronizes more devices than the block size (default 5000), because synchronization code clears the life cycle table before importing the block of life cycle states. So if more than one block is synchronized, the previous blocks are cleared before we can process them.
    Resolution:
    Update synchronization code to clear the life cycle table only after global synchronization completes successfully, instead of before each block of life cycle changes from Spectrum.
    (3.7.0, DE402580, 01271117)
  • Symptom:
    When Performance Center initiates a synchronization with a data source and the status is not success, the Last Polled On time will not update, leading to confusion to whether it's even being polled.
    Resolution:
    Updated Data Sources page to show the Last Polled On time whether it's successful or not. This will tell the user that Performance Center is still attempting to synchronize with the data source. See the state for whether it was successful or not.
    (3.7.0, DE402650, 01277928)
  • Symptom:
    The Group membership Extract Transform and Load (ETL) process on the Data Aggregator runs every minute by default and uses memory intensive data structures to perform this task. This can result in prolonged Data Aggregator garbage collection pauses and relatively low application throughput.
    Resolution:
    The Group ETL process was optimized for performance with respect to memory utilization. Previously, all intermediate data structures were maintained in memory while the data was written to the DR, and the data was written to the DR in bulk which further increased the memory footprint of this process. The process now clears all intermediate data structures prior to writing the output to the DR, which makes that memory eligible for garbage collection earlier, which significantly reduces the chance of application pauses, It also writes the data to the DR in batches which further reduces the memory footprint required and allows individual writes to complete faster allowing the allocated memory to be reclaimed in a more efficient manner.
    (3.7.0, DE402837, 01285065)
  • Symptom:
    Even though the IP address table has an IP address for an ifIndex, the IP address would not be displayed for an interface synced from Data Aggregator. Looking at the interface via data aggregator REST would also show no IP address for the interface. data aggregator logs would indicate a warning message about "Component Facet Values are not the correct length".
    Resolution:
    When validating cross-metric family attributes, handle when they may not have the same length. IP address are now associated to the interfaces correctly and synced to Performance Center.
    (3.7.0, DE403189, 01286655)
  • Symptom:
    On a fresh installation of Performance Center 3.6, the child items of devices will not be visible if running dashboards against the Default Domain group, but will be visible for any newly created IP Domain.
    Resolution:
    The item flags for the Default Domain item are set correctly on an upgrade or a fresh install of Performance Center 3.6.5 and higher. Once upgraded/installed, child components of devices in the Default Domain should be visible in dashboard views when Default Domain is the selected context group. (3.7.0, DE403284)
  • Symptom:
    Events may be missing in Events Display. When checking the Event Manager log files, you will see a NullPointerException with GroupPath.getFilteredParentIds(GroupPath.java:linenum) in stack trace.
    Resolution:
    Event Manager updated to check for null in groupPath. Added additional exception handling to ensure we process the events if groupPath has unexpected exception.
    (3.7.0, DE403441, 01274256,01300824)
  • Symptom:
    Snmp Poller processing tracking in data collector karaf log was incorrect due to being stored as static variables.
    Resolution:
    Each SnmpPoller object now has its own member variables so we can correctly print information to the data collector
    karaf.log
    file.
    (3.7.0, DE404037)
  • Symptom:
    Sync takes a long time to complete.
    Resolution:
    Add database index to field often used during sync.
    (3.7.0, DE404046)
  • Symptom:
    The data collector noSuchInstance suppression for multi-index components for QOS metric families is not working.
    Resolution:
    Refactored SNMP request functionality to correctly suppress noSuchInstance handling.
    (3.7.0, DE404217)
  • Symptom:
    For RIB queries with many items in return result, the translation of IDs in result could be slow.
    Resolution:
    Improved the ID translation in processing RIB result.
    (3.7.0, DE404638)
  • Symptom:
    CAPerformance Center takes a very long time and runs out of memory when building a Context page for a device that is in a very complex group hierarchy.
    Resolution:
    Optimized Context page group path finder to work in RAM instead of DB, and limited results to ten.
    (3.7.0, DE406338, 01317443)
  • Symptom:
    The data collector
    karaf.log
    file can get filled up with counter rollover messages.
    Resolution:
    Updated logging to no longer log to the
    karaf.log
    file. They are already being logged to the
    CounterRollover.log
    file.
    (3.7.0, DE406390)
  • Symptom:
    QueryBuilder blocks creation of a query with percentile and non percentile aggregation methods.
    Resolution:
    With this fix, QueryBuilder would allow to create a query with percentile and non percentile aggregation methods.
    (3.7.0, DE406569)
  • Symptom:
    Could not check the current active OData limiter settings used by the Data Aggregator.
    Resolution:
    Added function to output the current OData limiter settings when checked via Data Aggregator debug page.
    (3.7.0, DE406934)
  • Symptom:
    While processing query responses for certain CA Performance Center dashboard views, the Device Manager would use an inefficient MySQL query to translate identifiers between the data source and CA Performance Center leading to results being delayed to the user. The query, that includes the text "WHERE ii.LocalID IN (1, 2, 3)" in which the numbers are not quoted, would sometimes be present in the MySQL slow query log if that logging was enabled.
    Resolution:
    The query was modified to use appropriate quoting which allows MySQL to use a pre-existing index and significantly improves the performance of the query.
    (3.7.0, DE409328, 01327882)
  • Symptom:
    When more than 1 device from the Data Aggregator got consolidated in Performance Center, it caused both devices to have the same IP address in the data aggregator due to PUSH synchronization.
    Resolution:
    Updated Data Aggregator PUSH synchronization handler to no longer update the IP address of the device. Any device IP address update, via the device context page edit dialog, is now handled by a REST call from the Performance Center to the DA.
    (3.7.0, DE409712)
  • Symptom:
    Seeing No Data to Display on SD-Wan views when one of the sites is not reporting during time range.
    Resolution:
    Fixed issue when one of the SD sites is not reporting during time range that leads to No Data to Display on SD-Wan views.
    (3.7.0, DE411377, 01316086)