Skip to main content

Hello,

We are trying to set up our Postgres DB to use blocklevel backups. 
Our system is PostgreSQL 10.13 running on Linux RH7.

 

The backups run without an error. I can see all full and incremental (logs) that have been done over the last several days. After preparing the server (getting rid of the /main file and all log files) I started from the web interface a restore to current.

After about 15 seconds the Progress indicator goes from green to yellow (Shows 10% completed) and I get the the error message : Received failed message for job o258820], phase sRestore]

  • Event code 19:122

After that the only thing I can do is kill the Job amd go back to my older normal backup. CommVault can restore that without a problem.

Thanks for any help!

Douglas

Thanks for the post, Douglas!

Error 19:122 is a contextual message with the advice in the description depending on the issue.

Can you check the job details and see what the second <string> value says?

Received failed message for job [<Number>], phase [<String>].

We may need to see some log files depending on what the message says.


The other message is just: Received failed message for job 258820], phase Restore].

If you Need more info from a log just let me know which log file I should look into.

Thanks for the help! 


The log file is called PostgreSQLrestore.log.

Paste the messages in here (I’d advise you redact any client names with <ClientName>, ditto with IP addresses) and I’ll take a look!


@DSchat , following up to see if you were able to pull the error in the log file for review.

Thanks!


Log file name depends on the type of restore run. Is it restore from snap copy or primary copy or a volume level restore or any other type (table level or clone restore) ?


We found the solution. I was doing the restore from the web interface. You Need to run the restore from the CommVault Commcell Console. When the window for browse and restore Options open, you need to click on the Advanced Options tab and ther choose the corresponding media Agent. When you have the correct Media Agent the Block Level restore completes sucessfully.
The funny thing about this is that after that is set, the restore also runs from the web interface.


@DSchat , thanks for the follow up (I marked your own reply as the Best Answer).

If I had to explain why, my initial thought is that your first successful browse restored all of the necessary Index data (that was still there), so subsequent attempts worked from any interface.


Reply