Re: pg_xlog dir not getting swept

Thanks, after a few more hours of waiting, things started get cleaned up. Things are back in order now.

Niels Kristian Schjødt
Co-founder & Developer

E-Mail: nielskristian(at)autouncle(dot)com <mailto:nielskristian(at)autouncle(dot)com>
Mobile: +45 28 73 04 93

www.autouncle.com <http://www.autouncle.com/>
Follow us: Facebook <https://www.facebook.com/AutoUncle> | Google+ <https://plus.google.com/+AutoUncle> | LinkedIn <http://www.linkedin.com/company/autouncle> | Twitter <https://twitter.com/AutoUncle>
Get app for: iPhone & iPad <https://itunes.apple.com/en/app/autouncle/id533433816?mt=8> | Android <https://play.google.com/store/apps/details?id=com.autouncle.autouncle>

> Den 29. jun. 2016 kl. 21.19 skrev Jeff Janes <jeff(dot)janes(at)gmail(dot)com>:
>
> On Wed, Jun 29, 2016 at 3:00 AM, Niels Kristian Schjødt
> <nielskristian(at)autouncle(dot)com> wrote:
>> About a day ago, there seems to have been some trouble in the network of my
>> database (postgresql 9.3).
>>
>> I’m running my db with a streaming replication setup with wall shipping.
>>
>> I sync wal logs to a mounted networkdrive using archive_command = 'rsync -a
>> %p /mnt/wal_drive/wals/%f </dev/null’. Somehow this command was failing,
>> leading to my pg_xlog dir building up (590Gb). I rebooted the server, and
>> the archiving command seems to succeed now - however - After about an hour
>> of running, the pg_xlog drive has not decreased in size - I would have
>> expect that! I can see that lot’s of files get’s synced to the
>> /mnt/wal_drive/wals dir, but somehow the pg_xlog dir is not swept (yet)?
>> Will this happen automatically eventually, or do I need to do something
>> manually?
>
> Successfully archived files are only removed by the checkpointer. The
> logic is quite complex and it can be very frustrating trying to
> predict exactly when any given file will get removed. You might want
> to run a few manual checkpoints to see if that cleans it up. But turn
> on log_checkpoints and reload the configuration first.
>
> Cheers,
>
> Jeff