Re: Pgbackrest fails due after an ISP change

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(at)gmail(dot)com>
wrote:

>
>
> On Mon, Apr 21, 2025 at 9:03 AM KK CHN <kkchn(dot)in(at)gmail(dot)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.
>
> https://pgbackrest.org/command.html#command-check
>
>
> Cheers,
> Greg
>
> --
> Crunchy Data - https://www.crunchydata.com
> Enterprise Postgres Software Products & Tech Support
>
>