Thread: Postgres Wal Full

On 3 March 2022 08:46:45 CET, pgdba pgdba <postgresdba12@outlook.com> wrote:
>Hi ,
>
>I have a problem from pg_wal. I am using postgresql version 11 and taking backup and writing archive_command in
postgresql.confbut did not archive wal and my disk ise full from pg_wal. I research why is my wal is full and dont
foundanything could  you help me please ? 


Check if your archive_command works. Read the log.


--
2ndQuadrant - The PostgreSQL Support Company

At Thu, 3 Mar 2022 07:46:45 +0000, pgdba pgdba <postgresdba12@outlook.com> wrote in 
> I have a problem from pg_wal. I am using postgresql version 11 and taking backup and writing archive_command in
postgresql.confbut did not archive wal and my disk ise full from pg_wal. I research why is my wal is full and dont
foundanything could  you help me please ?
 

The immediate cause of the WAL retension is the archive failure.
PostgreSQL, while archiving is enabled, won't remove a WAL file until
it is successfully archived.

You will see the lines like the following in sergver log if archive
failure happened.

> 2022-03-03 17:15:21.314 JST LOG:  archive command failed with exit code 1

If the archive command emitted some error messages, it can be caught
in the server log together.  The follwoing error is emitted by the
archive_command "cp %p /archive/%f".

> cp: cannot create regular file '/archive/000000010000000000000001': No such file or directory

regards.

-- 
Kyotaro Horiguchi
NTT Open Source Software Center

On Thu, Mar 3, 2022 at 2:17 PM pgdba pgdba <postgresdba12@outlook.com> wrote:
>
> Hi, thanks your feedback but not have same error in my log file. I am using streaming replication and have 1 master
and2 slave nodes in async. master node different size file wal from slave nodes such as;
 
> [root@master1 pg_wal]# ls | wc -l
> 6400
> [root@slave2 pg_wal]# ls | wc -l
> 332
> [root@slave3 pg_wal]#  ls | wc -l
> 344
>
> wal_keep_segments = 512
> max_wal_size = 2GB
> min_wal_size = 1GB
> data size : 290G
>  cpu:8
> ram:16

Few things to check when WAL file grows: 1) archive failures 2)
inactive replication slots (use select * from pg_replication_slots;)
3) infrequent checkpoints (use select * from pg_control_checkpoint; to
know the last checkpoint time)4) high write (WAL-generating)
workloads.

Regards,
Bharath Rupireddy.