Thread: Pgbackrest fails due after an ISP change

Pgbackrest fails due after an ISP change

From
KK CHN
Date:
List,

I am running  a pgbackrest(2.52.1) backup setup for a postgres(16) server(RHEL9.4)  to a repo server RHEL 9.4..   

The pgbackrest was working fine and the remote repo server got regular backups through cron scheduler. 

Suddenly there was an ISP change at our  end,  the VPN tunnel  between DB server and Repo server was lost for  24 hours due to VPN tunnel crash.  later I restored the VPN tunnel

After that  the cron scheduler does not perform the pgbackrest backup. It fails!!!

So I performed a full backup to overcome this ..  But it fails  with the cron command (  38 14 * * 1 pgbackrest --type=full  --stanza=My_Repo backup) 


ERROR: [082]: WAL segment 000000010000022000000038 was not archived before the 60000ms timeout
                                    HINT: check the archive_command to ensure that all options are correct (especially --stanza).
                                    HINT: check the PostgreSQL server log for errors.
                                    HINT: run the 'start' command if the stanza was previously stopped.
                                    --------------------------------------------------------------------
                                    If SUBMITTING AN ISSUE please provide the following information:

                                    version: 2.52.1
                                    command: backup
                                    options: --delta --exec-id=492671-6f2068fc --log-level-console=info --log-level-file=debug --pg1-host=10.x.y.202 --pg1-host-user=enterprisedb --pg1-path=/data/edb/as16/data --pg-version-force=16 --process-max=5 --repo1-block --repo1-bundle --repo1-cipher-pass=<redacted> --repo1-cipher-type=aes-256-cbc --repo1-path=/DBBACKUP --repo1-retention-diff=3 --repo1-retention-full=3 --stanza=KTA_Repo --start-fast --type=full
2025-04-21 14:39:07.335 P00   INFO: backup command end: aborted with exception [082]
2025-04-21 14:39:07.335 P00  DEBUG:     command/exit::exitSafe: => 82
2025-04-21 14:39:07.936 P00  DEBUG:     main::main: => 82

How can I make the full backup command not to check the  WAL was archived or not  to the repo server for atleast once ?

Any hint  appreciated .. 

Thanks
Krishane


Re: Pgbackrest fails due after an ISP change

From
Greg Sabino Mullane
Date:


On Mon, Apr 21, 2025 at 9:03 AM KK CHN <kkchn.in@gmail.com> wrote:

ERROR: [082]: WAL segment 000000010000022000000038 was not archived before the 60000ms timeout
... 
How can I make the full backup command not to check the  WAL was archived or not  to the repo server for atleast once ?

You cannot. WAL is an integral part of the backups. You should run the pgbackrest "check" command and look at your Postgres logs to figure out why WAL files are not being archived. Once that is solved, try the backup again.



Cheers,
Greg

--
Enterprise Postgres Software Products & Tech Support

Re: Pgbackrest fails due after an ISP change

From
KK CHN
Date:
Hi List, 

Thanks for the clarification.  

 If I want to start a fresh backup to the same repo server ( the one which fails the pgbackrest backup due to the file system full outage) what all are  the steps I have to perform on the Repo Server and the DB server ?

AFAIK,    These all are the steps  required , correct me if I'm wrong.

1.  Stop the   stanza on both repo server as well as on DB server      ( $sudo -u  myuser  pgbackrest --stanza=My_RepoX  --force stop) 
2.  Then delete the stanza from either Repo Server or DB server (  by the delete stanza option ) Is the deletion of the stanza needed to be performed only at the Repo server is enough ?
3.  After deletion of the stanza,   delete the folders in the repo server   archive, backup respectively by ( rm -f   archive    , rm -rf backup   folders  for My_RepoX ?)
4. Then execute the same  stanza create command  from the reposerver  ?  ( $ sudo -u myuser pgbackrest  --stanza=My_RepoX  --log-level-console=info stanza-create)?
5.  Then Check the stanza creation and config all perfect by  (  sudo -u myuser pgbackrest --stanza=My_RepoX  --log-level-console=info check ) 
6.   If the step 5 success,   schedule the full backup from the cron 


I hope  this will help me to take backup afresh in case an existing pgbackrest backup  fails ( whatever the stupid reason what happened as in my case ) . 

( In my case pgbackrest fails due to the scenario I faced, I mean due to the file system full  and didn't notice it for 24 hours, even after made space of the partition /root which got full,  later on the pgbackrest backups from cron scheduler unable to complete the backup)


   If anything is wrong or not sufficient as in the 6 steps AFAIK (As this a live Repo Server for other DB Clusters too, more than 10 nos, nothing should go wrong) ,   please correct/guide me   

Thank you,
Krishane



On Mon, Apr 21, 2025 at 9:49 PM Greg Sabino Mullane <htamfids@gmail.com> wrote:


On Mon, Apr 21, 2025 at 9:03 AM KK CHN <kkchn.in@gmail.com> wrote:

ERROR: [082]: WAL segment 000000010000022000000038 was not archived before the 60000ms timeout
... 
How can I make the full backup command not to check the  WAL was archived or not  to the repo server for atleast once ?

You cannot. WAL is an integral part of the backups. You should run the pgbackrest "check" command and look at your Postgres logs to figure out why WAL files are not being archived. Once that is solved, try the backup again.



Cheers,
Greg

--
Enterprise Postgres Software Products & Tech Support