From: | Jehan-Guillaume de Rorthais <ioguix(at)free(dot)fr> |
---|---|
To: | Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> |
Cc: | alexanderfelipewo <alexanderfelipewo(at)gmail(dot)com>, pgsql-general(at)postgresql(dot)org |
Subject: | Re: pgbouncer increase pool_size, reload does not work |
Date: | 2017-02-03 08:48:06 |
Message-ID: | 20170203094806.6889e652@firost |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Thu, 2 Feb 2017 16:12:06 -0800
Adrian Klaver <adrian(dot)klaver(at)aklaver(dot)com> wrote:
> On 02/02/2017 03:27 AM, alexanderfelipewo wrote:
> > hello!
> >
> > i have a question for pgbouncer in case someone has faced this and there is
> > a solution available.
> >
> > i have a db with pgbouncer where the pool_size is set to 50 (i ve noticed
> > this using both session/transaction modes). Sometimes there are waiting
> > sessions b/c more than 50 want to execute sth at once and usually each query
> > is not so fast. So let's say i want to increase the pool_size from 50 to 70.
> > I change the config file and then login to pgbouncer where i run a
> > 'reload;'. Now when i check the 'show config' i see the new value in the
> > 'default_pool_size' parameter.
> >
> > BUT
> >
> > the problem is that the amount of servers is still 50 and does not go up to
> > 70. At the same time there are clients in a 'waiting' state. If i do a
> > restart then it will work and the servers will be able to go up to 70.
> >
> > I ve tried these in different databases and it doesnt happen all the time
> > but it does happen pretty often.
> > Has anyone else seen this? if so, is there a solution (except restarting) or
> > explanation?
>
> Would this work?:
>
> https://pgbouncer.github.io/faq.html#how-to-upgrade-pgbouncer-without-dropping-connections
Maybe, but I couldn't give it a try as my pgbouncer is managed by Systemd.
From | Date | Subject | |
---|---|---|---|
Next Message | JP Jacoupy | 2017-02-03 09:18:40 | Re: Synchronous Commit, WAL archiving and statement_timeout |
Previous Message | Andreas Joseph Krogh | 2017-02-03 08:29:56 | Re: RUM-index and support for storing BIGINT as part of a tsvector+timestamp |