Solved

PostgreSQL Block-Level Backup Snap Job is skipped


Userlevel 3
Badge +7

Hello,

 

Our Customer has PostgreSQL systems. We are backing up some of these systems with block-level. I started seeing errors in incoming reports;

Error Code: [80:123]
Description: Snap Job [<Job ID>], Storage Policy [<Policy Name>] is skipped for backup copy as the job is not picked.
Source: <commserve>, Process: JobManager

 

We have a storage policy with snapshot copy and spool copy is enabled. In Job Selection Rule all backups selected.

 

I couldn't find anything useful in the logs.

 

Can anyone help me with this?

 

Commvault Version: 11.20.27

Commserve is a virtual machine Server 2019 installed.

Media Agent is a physical machine Server 2019 installed.

PostgreSQL installed on RedHat 7.7 and PostgreSQL version is 12.4.

icon

Best answer by Meera 4 March 2021, 12:45

View original

10 replies

Userlevel 3
Badge +7

Hi Ibrahim,

PostgreSQL log backups are regular streaming backups only and not snap backups. They do not require a backup copy run. Is there any schedule present that could be causing this? Is it possible to share a screenshot of schedules associated to the instance and details of backup copy schedule?

 

Regards,

Meera

Userlevel 3
Badge +7

Hi Meera,

Thank you for reply.

I am attaching screenshots.

 

Regards,

İbrahim

Userlevel 3
Badge +7

Hi Ibrahim,

Please edit schedule named “LOG” and select advanced and deselect the option to “Create Backup Copy immediately” as this is not needed for streaming backup. Since this option is selected, it is launching backup copy for log backups and since log jobs do not require backup copy it is marked “Failed to Start”. I will check with team to prevent invoking job in this case. But meanwhile, editing schedule should help here. 

 

Also, why log backup took 5 hours? Was there so much logs to backup?

 

Regards,

Meera

Userlevel 3
Badge +7

Hi Ibrahim,

Please edit schedule named “LOG” and select advanced and deselect the option to “Create Backup Copy immediately” as this is not needed for streaming backup. Since this option is selected, it is launching backup copy for log backups and since log jobs do not require backup copy it is marked “Failed to Start”. I will check with team to prevent invoking job in this case. But meanwhile, editing schedule should help here. 

 

Also, why log backup took 5 hours? Was there so much logs to backup?

 

Regards,

Meera

Hi Meera,

I edited schedule which name is LOG and i deselected “Create Backup Copy immediately”.

But we have another PostgreSQL machine which is backs up with block-level too. This machine has only Full and INC data schedules. This morning i saw error on this too. Do i need deselect “Create Backup Copy immediately” in INC Data too?

Log backup size around 1TB and Throughput around 200GB/H. Screenshot attached.

Userlevel 3
Badge +7

Hi Ibrahim,

Client 1- Are we having old logs in the archive log directory without any cleanup? If yes, is there any reason for not deleting logs which are backed up? 

If deleting old logs is fine, please enable ”Archive Delete” option in instance properties, this will cleanup backed up logs. https://documentation.commvault.com/commvault/v11/article?p=21520.htm 

Client 2 - Incremental data jobs are snap backups and enabling backup copy immediately for those jobs is fine. What is the job type for failed to start job? Is it showing same JPR as Log Only job (Snap Job is skipped for backup copy as the job is not picked) Or please send a screenshot similar to the first one you send, from backup history.

 

Regards,

Meera

Userlevel 3
Badge +7

Hi Ibrahim,

Client 1- Are we having old logs in the archive log directory without any cleanup? If yes, is there any reason for not deleting logs which are backed up? 

If deleting old logs is fine, please enable ”Archive Delete” option in instance properties, this will cleanup backed up logs. https://documentation.commvault.com/commvault/v11/article?p=21520.htm 

Client 2 - Incremental data jobs are snap backups and enabling backup copy immediately for those jobs is fine. What is the job type for failed to start job? Is it showing same JPR as Log Only job (Snap Job is skipped for backup copy as the job is not picked) Or please send a screenshot similar to the first one you send, from backup history.

 

Regards,

Meera

Hi Meera,

1- I talked this situation with Customer. They said its a test site its not important. I’ll look at it.

2- I’m seeing this error (Snap Job is skipped for backup copy as the job is not picked) only incremental data and full backups. Log backups runs every 30min and no error. I attached screenshots.

Thank you.

Userlevel 3
Badge +7

Hi Ibrahim,

If you have snap engine type as native (subclient properties), please disable Create backup copy immediately option. If it is hardware engine, please escalate the issue. We should not be triggering backup copy for log backup for PostgreSQL agent. 

 

Regards,

Meera 

 

Userlevel 3
Badge +7

Hi Ibrahim,

If you have snap engine type as native (subclient properties), please disable Create backup copy immediately option. If it i.s hardware engine, please escalate the issue. We should not be triggering backup copy for log backup for PostgreSQL agent. 

 

Regards,

Meera 

 

Hi Meera,

I am using snap engine as native. In documentation it says 

So i need to choose Create Backup Copy immediately option, is it right? I have 3 schedule. One for Full, one for Incremental Data and last one for Incremental Log backups. So i enabled Create Backup Copy immediately option for Full and Incremental Data. I didn’t enable it for Incremental Log.

Documantation says “After the full backup job is complete, log backup job starts automatically.” So when Full or Incremental Data backups finished Transaction Log runs.

This is what i see; When the Full or INC Data backups is finished the log backup is starts. Meanwhile, backup copy workflow is running for Full or INC Data backups. Two processes are working at the same time. And it says it skipped when log backup is finished. It completes backup copy without any problems for Full or INC Data.

I tried to disable or enable Create Backup Copy immediately. The situation is the same for both.

Is it possible to disable Create Backup Copy immediately option for Transaction Log when is enabled for Full or INC Data. I mean for log backup which is automatically running after Full or INC Data backup.

One more thing. I am using 1 storage policy for all PostgreSQL systems (for normal backups, block-level backups, log backups). Do i need different storage policy for log backups?

Thank you.

Userlevel 3
Badge +7

Hi Ibrahim,

When snap engine is native, backup copy is triggered immediately automatically. (This is there for sp20 also) Since you have engine type as native, Create backup copy immediately option is not required to run backup copy.

 

Having the option enabled doesn't cause any harm except for triggering backup copy for log job and mark it failed to start. We need to check that issue and avoid it. The job will be seen only when checking backup job history. It doesn't connect to client or do any operation. As there is no snapshot for log backup, backup copy is not needed and so the job status goes to failed to start status immediately after it is created. 

 

But you mentioned that the issue is there after disabling Create Backup Copy immediately option also.
That is not seen in the lab setup. In the attached screenshot, 102130777 is the problematic job. This was not triggered for the next log backup after disabling Create backup copy immediately option. If disabling option is not helping, we would need more details to check it further, please escalate. If disabling option helps, please create a CMR to track the issue.

 

There is no requirement that log backup should use separate storage policy.
It can be same or different - both are supported.

Userlevel 3
Badge +7

Hi Meera,

Ok I think I understand now.

I disabled Create backup copy immediately option on test PostgreSQL system and it worked. It did not give error.

I disabled Create backup copy immediately option on prod PostgreSQL system but it gave error again.

But then I remembered, I entered nDoNotBackupAllLogs additional setting to test server.

https://documentation.commvault.com/additionalsetting/details?name=nDoNotBackupAllLogs

The customer doesn’t want to prune logs.

I entered this additional setting on prod server and it worked.

I think log backups take a long time because it did not prune. It gives error because backup copy is completed at that time.

I attached screenshots. You can ignore 219557. Its a different job.

 

Thank you for your helps and patience Meera, I appreciate.

Reply