New update available - Opslogix Oracle Management Pack (V.26.11.4544.0)

We are happy to announce that a new release of our Opslogix Oracle Management Pack for SCOM (V.26.11.4544.0) is now available.

You can download the latest version from our Customer Portal.

This release adds four monitors for problems that used to go unnoticed until it was too late: a failed RMAN backup, a monitoring account whose password is about to expire, a pluggable database that stays closed after a restart, and an archive gap that only the Data Guard primary can see. Alert log alerts are now quieter and get the right severity, and a mounted physical standby database is monitored much more completely.

Before you upgrade

• Monitoring accounts with minimal privileges need three more grants: SELECT on sys.gv_$containers, sys.v_$archive_dest_status and sys.gv_$logstdby_process.
• After the import, run Test Connection in the Oracle IMP Configuration Dashboard to update the Oracle collector, and then restart the Microsoft Monitoring Agent on the resource pool members.

The Upgrade Guide lists all steps and what changes after the upgrade.

New updates:

• Added: Monitor 'Failed Backups': raises a critical alert when the latest run of an RMAN backup type (for example DB FULL, DB INCR, ARCHIVELOG) is FAILED or COMPLETED WITH ERRORS, and closes the alert after a later run of that type succeeded. Before, a failed backup was only reported when the backup age monitors found the last successful backup too old, which could be a day later. Runs every 15 minutes on primary and standby databases; the LookbackHours override (default 48) sets how far back runs count. See How backup monitoring works.
• Added: Monitor 'Monitoring Account Password Expiry': raises a warning alert when the password of the monitoring account expires in less than 14 days, and a critical alert when it has expired and the account is in its grace period. Oracle's DEFAULT profile expires passwords after 180 days, after which monitoring of the instance stops. Runs on primary databases and needs no extra grant. SYSDBA connections stay Healthy. See Troubleshooting: The monitoring account password expires.
• Added: Monitor 'Pluggable Databases Open': checks from the CDB root that every pluggable database of the local instance is open READ WRITE. Critical when a PDB is MOUNTED or open READ ONLY, Warning when it is open RESTRICTED; the alert names the PDBs. A PDB that stays MOUNTED after a restart is not discovered, so before no monitor reported it. Use the ExcludedPDBs override for PDBs that are closed or read-only on purpose.
• Added: Data Guard monitor 'Standby Archive Gap (primary)': raises a critical alert when the primary reports an archive gap for a standby destination. The existing archive gap monitor on the standby cannot see a gap when the standby receives nothing from the primary.
• Added: Alert log rule: CriticalOraNumbers override to choose which ORA numbers raise a Critical alert.
• Changed: Instance CPUCount property now comes from the CPU_COUNT parameter instead of DBA_CPU_USAGE_STATISTICS. This is the same value unless CPU_COUNT is set lower than the number of CPUs, for example for instance caging.
• Changed: Oracle collector: ORA-01219 (database not open) is logged as verbose instead of as a warning.
• Changed: Knowledge articles: links to the matching Oracle Database 19c documentation for 97 discoveries, rules and monitors, consistent formatting in the Operations console, Causes and Resolutions for the critical ORA numbers of the alert log rule, and the backup monitors explain that they stay Healthy when no RMAN backup exists, by design.
• Fixed: Alert severity now follows the monitor state: a Critical state raises a Critical alert, a Warning state a Warning alert. Before, the seven backup monitors and the Session utilization, Process utilization and Invalid Indexes monitors always raised a Warning alert, also when Critical. An override on the alert severity of these monitors still applies; remove it to get the new behavior.
• Fixed: Alert log rule: every ORA- entry raised a Critical alert. ORA numbers in the CriticalOraNumbers list (default 600, 7445, 4031, 1578, 257, 19809, 27102, 1157, 29740) now raise a Critical alert, all others a Warning.
• Fixed: Alert log rule: repeated ORA- entries were never suppressed. There is now one open alert per ORA number per instance, and repeats increase the Repeat Count.
• Fixed: Data Guard transport lag and apply lag reported 0 seconds when the standby received no data from the primary. They now report the time since the standby last received data, so a broken redo transport raises an alert.
• Fixed: Data Guard logical recovery monitor never detected a stopped SQL Apply on a logical standby database. It now checks GV$LOGSTDBY_PROCESS.
• Fixed: Mounted physical standby: performance collection failed (ORA-01219, then ORA-01476), so the standby collected no performance data. A mounted standby reads no buffers, so its buffer hit ratio is reported as 100%.
• Fixed: Mounted physical standby: the CPU count, physical database size and clustered discoveries and the Database size performance rule failed with ORA-01219. They now read fixed views, which are available on a mounted database.
• Fixed: Instance IsClusteredDatabase property was never discovered on any database. It now reads the CLUSTER_DATABASE parameter and shows Yes or No.
• Fixed: Top 10 SQL tasks failed with 'PostDataItems failed' when the text of a statement contained a NUL character, which several internal Oracle statements do.
• Fixed: Tasks showed Success with 'No SQL output.' when their query failed, for example on a mounted standby. The task output now shows the Oracle error, and 'No SQL output.' only means that the query returned no rows.
• Fixed: Instance Up Since property showed the month instead of the minutes. It now uses the format YYYY-MM-DD HH24:MI:SS.
• Fixed: Instance Multitenancy property: 'PUGGABLE DB' is now 'PLUGGABLE DB' and 'Could not be dectected' is now 'Could not be detected'. Update your own groups or views that filter on the old values.
• Fixed: Instance Process Utilization monitor labelled its value as 'Session Utilization Pct' in the monitor state context; it is now 'Process Utilization Pct'.
• Fixed: Monitoring account script: the minimal-rights grant for GV$CONTAINERS named a view that does not exist, so the account could not read the view that the instance discovery uses to find pluggable databases.
• Fixed: Typos in the display strings of the 10g alert log rule and the Session utilization, Process utilization and Invalid Indexes monitors.

Haven't had the chance to try the Opslogix Oracle Management Pack yet?

Start your 30-day free and non-binding trial below!

Are you unable to find the new version in the portal? Contact us to try the newest version or to get assistance updating your license key.