← Back to team overview

maria-discuss team mailing list archive

Re: MariaDB server horribly slow on start

 

Thanks, I'll keep that in mind indeed. I was very surprised by the system limits (using mariabackup) so I couldn't imaging that being such an issue.

After a general failurea few days ago I was left with one node alone and starting a second node produced the same issue while only one node was used for reading and writing. So writing to multiple nodes might be another issue I'll have to deal with later.

If I can solve that restart issue, that'd be removing a huge problem which puts down all our services entirely for 30 minutes when it happens.

So again, thanks a lot for your precious help, hopefully that setting is getting us in the right direction.

-----Message d'origine-----
De : Gordan Bobic <gordan.bobic@xxxxxxxxx> 
Envoyé : jeudi 28 juillet 2022 10:41
À : Cédric Counotte <cedric.counotte@xxxxxxxxxx>
Cc : jocelyn fournier <jocelyn.fournier@xxxxxxxxx>; Marko Mäkelä <marko.makela@xxxxxxxxxxx>; Mailing-List mariadb <maria-discuss@xxxxxxxxxxxxxxxxxxx>; Pierre LAFON <pierre.lafon@xxxxxxxxxx>
Objet : Re: [Maria-discuss] MariaDB server horribly slow on start

On Thu, Jul 28, 2022 at 11:24 AM Cédric Counotte <cedric.counotte@xxxxxxxxxx> wrote:
>
> I've prepared all servers with that new setting, and this (is it ok or should I set it to 1048576 as well?):
>         table_open_cache = 65536

No, remember that LimitNOFILE has to fit at least the sum of all of the settings I mentioned, and still leave plenty of headroom.
If you had PMM up and running you could see at a glance whether your table_open_cache is big enough.

> I'll do the server restart this evening to avoid creating problems during the day.
>
> I did try to restart the backup cluster (2 nodes, one slave of the main cluster) and it didn't seem to slowdown the slave as it used to, so that might be the solution (or part of)!
>
> Thanks a lot for your time, will keep you posted later today when I restart a node of the main cluster.

IMO you are probably experiencing bigger problems from writing to multiple nodes at the same time.

References