[pgpool-general: 5004] Re: Avoiding downtime when pgpool changes require a restart

[Muhammad Usama]

On Thu, Sep 15, 2016 at 9:43 PM, Jacobo García López de Araujo <
jacobo.garcia at gmail.com> wrote:

> I believe this is currently a bug, I'd like to know better before I fill a
> bug against PgPool bug tracker, I'm afraid that I'm missing something, but
> after one more day of tests, I have been unsuccessful of restarting 2
> PgPpool watchdog without incurring in a few seconds of downtime.
>
> I'll be grateful for any help or information about this issue.
>
> Many thanks,
>
> Jacobo García.
>
>
>
Hi

Your use case is valid and pgpool-II should not produce a FATAL error if
the configurations of nodes differs. I have pushed the fix for that in
pgpool-II 3.5 and master branches, You can try building from the source
code to check if your problem is fixed

https://git.postgresql.org/gitweb/?p=pgpool2.git;a=commitdiff;h=a38fa0910f94dfc5314fe34bd8ad86dc7dfb594e

Thanks
Best regards
Muhammad Usama

>
>
> On Wed, Sep 14, 2016 at 4:12 PM Jacobo García López de Araujo <
> jacobo.garcia at gmail.com> wrote:
>
>> Hello,
>>
>> I am trying to set load_balance_mode =  off setting in one testing 2
>> nodes cluster. The option is currently set to on. It is a setting that
>> requires a full restart in order to be changed.
>>
>> I haven't found a solution that does not provoke downtime on my setup.
>>
>> If I restart pgpool on the master node the watchdog is failed over to
>> the secondary, then master will refuse to join the cluster with pgpool
>> logs spitting the following error:
>>
>> FATAL:  configuration error. The configurations on master node is
>> different
>>
>> Then the master will shut down, every time I start the now old master it
>> will refuse to join the cluster because settings are different. In this
>> situation I just have one node running pgpool so if I restart this node
>> it will stop accepting connections through the virtual IP and downtime
>> will occur.
>>
>> The other strategy I tried is restart pgpool on the secondary node
>> first. In this case I also got the same error, and the secondary node
>> refuses to join the cluster too.
>>
>> I'd like to know what is the ideal procedure in order to change one of
>> those settings without having downtime.
>>
>> Many thanks for your time.
>>
>> Jacobo García.
>>
>>
>>
>> --
>> Jacobo García López de Araujo.
>>
> --
> Jacobo García López de Araujo.
>
> _______________________________________________
> pgpool-general mailing list
> pgpool-general at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-general
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20160920/c745f3a5/attachment.html>