Hello, Am Dienstag, 11. Juni 2019, 18:13:34 CEST schrieb Martin Caj:
I did a quick check today and the problem is on mirrordb1 with disk space
There is a disk /dev/vdb mount as /var/lib/pgsql. The disk is 100GB big and its full.
Right, that's what the monitoring also says since Friday.
on the PG-slave
the disk /dev/vdb (the same size) has 15 GB free disk space..
... and, according to the monitoring, wastes another 12 GB: POSTGRES_BLOAT CRITICAL: DB "mb_opensuse2" (db mb_opensuse2) table public.filearr rows:29279714 pages:2273999 shouldbe:587601 (3.9X) wasted size:13814972416 (12 GB) so after fixing the full disk on the master, you should probably run the cleanup routines (vacuum?) on the slave.
anyway I spoke with my colleague and they will extend the data disk Tomorrow. (if there is enough space felt on the storage).
We had a similar situation (disk full) a while ago, and _IIRC_ the problem was caused by old replication logs not being deleted. Maybe you should check for those logs (and ensure they get dropped automatically once old enough) instead of blindly adding more disk space just to delay the problem to next week ;-) As a sidenote: The mirrordb*.infra.o.o machines are not in salt and not connected to the FreeIPA accounts, which means I can't login there. (That's not as bad as it sounds because I don't know much about Postgres and wouldn't touch it anyway ;-)
I'm sorry for this, but since Theo left and Thorsten leaving now we have no replacement....
Managing the infra for openSUSE is harder and harder ....
It's not really harder, we "only" lack some people ;-) Regards, Christian Boltz --
Surely a nice initiative - but maybe over-engineered? How do you define over-engineered? I just spent 1-2h on it, so there cannot be much engineering in it. [> Dominique Leuenberger and Bernhard M. Wiedemann in opensuse-factory]
-- To unsubscribe, e-mail: heroes+unsubscribe@opensuse.org To contact the owner, e-mail: heroes+owner@opensuse.org