Re: postgrsql 9.5: Old WAL files remain in secondary `pg_xlog`

Hi

On Wed, Jun 5, 2024 at 5:10 AM Mohan NBSPS <mohan(dot)nbs(dot)ont(at)gmail(dot)com> wrote:

> thank you,
> we are not using slots due to storage related issues that could occur if
> the network is broken.
>
Slots are very useful and it will ensure that the Standby will not lag
behind the master and master will retain the WAL Files if standy did not
apply them yet, the storage is consideration but only incase of network
failure but this wont be a very big consideration.

Regards
Kashif Zeeshan
Bitnine Global

>
> sorry, my question was incomplete.
> I see few very old (over 2 years) WAL files remain in the secondary
> (primary is clean with only
> new WAL files).
>
> On Mon, Jun 3, 2024 at 9:33 PM vignesh kumar <
> vigneshkumar(dot)venugopal(at)outlook(dot)com> wrote:
>
>> You can use slot based replication, so that postgres knows which WAL FILE
>> to retain and delete. however, if you're not concerned about storage, but
>> then the replica go for WAL storage or WAL hub kind of solutions..
>>
>> Get Outlook for Android <https://aka.ms/AAb9ysg>
>> ------------------------------
>> *From:* Kashif Zeeshan <kashi(dot)zeeshan(at)gmail(dot)com>
>> *Sent:* Sunday, June 2, 2024 5:17:57 PM
>> *To:* mohan(dot)nbs(dot)ont(at)gmail(dot)com <mohan(dot)nbs(dot)ont(at)gmail(dot)com>
>> *Cc:* pgsql-admin(at)lists(dot)postgresql(dot)org <pgsql-admin(at)lists(dot)postgresql(dot)org>
>> *Subject:* Re: postgrsql 9.5: Old WAL files remain in secondary `pg_xlog`
>>
>> Hi Mohsn
>>
>> There can be multiple reasons for that.
>>
>> 1. If the archiving process on the standby is not working correctly, old
>> WAL files will not be removed.
>> 2. Ensure that the configuration settings for WAL management are
>> correctly set on both primary and standby
>> 3. If checkpoints are not happening frequently enough, WAL files will
>> accumulate.
>>
>> To diagnose the issue
>>
>> 1. Ensure that the replication between the primary and the standby is
>> working correctly.
>>
>> On Primary Server
>> SELECT client_addr, state, sent_lsn, write_lsn, flush_lsn, replay_lsn
>> FROM pg_stat_replication;
>>
>> On Standby Server
>> SELECT pg_last_wal_receive_lsn(), pg_last_wal_replay_lsn();
>>
>>
>> 2. Check if the archiving process on the standby is working properly.
>> Look for any errors or warnings in the PostgreSQL logs related to WAL
>> archiving.
>> 3. Ensure that the standby server has the correct restore command
>> configured to fetch the archived WAL files.
>>
>>
>> Regards
>> Kashif Zeeshan
>> Bitnine Global
>>
>>
>>
>> On Sat, Jun 1, 2024 at 4:01 AM Mohan NBSPS <mohan(dot)nbs(dot)ont(at)gmail(dot)com>
>> wrote:
>>
>> Hello Community,
>>
>> Although the WAL retaining settings are default (max size 1GB),
>> `wal_keep_segments` is 25, checkpoint target is 2.5 minutes(default).
>> I noticed a few WAL files in secondary nodes (streaming replication)
>> `pg_xlog`
>> folder.
>>
>> What could be the reason why postgresql did not delete these ?
>>
>> Thank you
>> Regards
>> Mohan
>>
>>