Let me try to explain about the HANA persistent jobs.
HANA database continuously initiates log backups. They can be as quick as a job in a few seconds or can be one job in an hour. It also backs up the catalog periodically (even if there is no activity), typically once in 15 minutes (this is also backed up by persistent log backup job).
When we receive the first log backup request from HANA, we will initiate a persistent log backup job and reserve a resource. Even if the log backup completes from HANA end, we will keep the persistent job alive and hold the resource. We keep this persistent job alive for 6 hours (configurable).
Three cases when we complete the CV persistent log backup job:
- When we receive the first log backup request after the 6 hours duration, we will complete the existing persistent job, release the resource, and start a new persistent job. So, in effect, the job may run for a few minutes over 6 hours.
- If there is any failure within these 6 hours, we will complete the job immediately. A new job will be started for the next log backup attempt from HANA. In this case, the persistent log backup job duration may be less than 6 hours.
- If HANA is not sending any log backup requests for 60 minutes. After an idle time of 60 minutes, we will complete the job, release resource and will not start a new persistent log backup job until HANA requests for a new log backup.
We map the HANA log backup jobs to CV persistent jobId based on the HANA DB service name. All the log jobs initiated by a HANA service during the 6 hours duration will be backed up by the same persistent job.
There are multiple HANA services per instance doing the log backups, that’s why we may see multiple persistent jobs running per Commvault instance at the same time.