[pgpool-general: 7495] Re: pgpool2 in kubernetes dies

[Nikolay Mastilo (HM)]

Thank you for the reply.

client_idle_limit is already set to 0, but clients still lose connection.

Can there be an impact that pgpool is running in k8s ?


Regards,
  Nikolay

12.04.2021 8:53, Umar Hayat пишет:
> Can this be following configuration?
>
> client_idle_limit = 0
>                                    # Client is disconnected after 
> being idle for that many seconds
>                                    # (even inside an explicit 
> transactions!)
>                                    # 0 means no disconnection
>
> Regards
> Umar Hayat
>
> On Mon, Apr 12, 2021 at 2:26 AM Nikolay Mastilo (HM) 
> <nikm71 at hotmail.com <mailto:nikm71 at hotmail.com>> wrote:
>
>     Please help to understand why frontend clients  lose connection to
>     pgpool after some idle and how can it be overcome.
>
>     Regards,
>       Nikolay
>
>     09.04.2021 12:36, Nikolay Mastilo (HM) пишет:
>>
>>     our clients are microservices running on k8s. and if they are
>>     connected directly to database, the connection resume successfully.
>>
>>     and if they are connected to pgpool, they cannot resume
>>     connection after some time.
>>
>>     some clients outside of kubernetes (for ex. pgAdmin) *cannot
>>     *resume connection after some time too. but the DBeaver resumes
>>     connection successfully.
>>
>>     magic
>>
>>     09.04.2021 12:01, Anton Melser пишет:
>>>
>>>         And we have such issue - after approx 2 hours of idle our
>>>         client service cannot resume connection with pgpool, we have
>>>         to restart client service only that doesn't acceptable for us.
>>>
>>>         Could you tell why doesn't  pgpool accept requests after 2
>>>         hours of idle and what we should do?
>>>
>>>     And you definitely don't have this with normal pgpool outside of
>>>     Kubernetes and these clients? I am pretty sure I have python
>>>     clients (inside the cluster) that sometimes aren't doing
>>>     anything for multiple hours and they stay connected fine. Your
>>>     clients are also inside your cluster, right?
>>>
>     _______________________________________________
>     pgpool-general mailing list
>     pgpool-general at pgpool.net <mailto:pgpool-general at pgpool.net>
>     http://www.pgpool.net/mailman/listinfo/pgpool-general
>     <http://www.pgpool.net/mailman/listinfo/pgpool-general>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pgpool.net/pipermail/pgpool-general/attachments/20210412/eea2c0e3/attachment-0001.htm>