[pgpool-general: 1052] Re: Issue with watchdog after failover

Lonni J Friedman netllama at gmail.com
Fri Sep 28 23:02:38 JST 2012


I've seen similar behavior, except that I ended up with both hosts
bringing up the delegate_IP simultaneously.  I sent a message to the
list a few days ago, but no one has replied.

On Fri, Sep 28, 2012 at 3:28 AM, Tim Verhoeven
<tim.verhoeven.be at gmail.com> wrote:
> On Fri, Sep 28, 2012 at 10:28 AM, Tim Verhoeven
> <tim.verhoeven.be at gmail.com> wrote:
>>
>> But I have a problem after the failover of pgpool. See the log below
>>
>> Sep 27 16:11:56 test01 pgpool[17847]: wd_init: start watchdog
>> Sep 27 16:11:56 test01 pgpool[17847]: pgpool-II successfully started.
>> version 3.2.0 (namameboshi)
>> Sep 27 16:11:56 test01 pgpool[17847]: find_primary_node: primary node id is 0
>> Sep 27 16:11:57 test01 pgpool[17854]: watchdog: lifecheck started
>> Sep 27 16:14:17 test01 pgpool[17853]: wd_escalation: escalated to master pgpool
>> Sep 27 16:14:19 test01 pgpool[17853]: wd_escalation:  escaleted to
>> delegate_IP holder
>> Sep 27 16:14:48 test01 pgpool[17854]: wd_lifecheck: lifecheck failed 3
>> times. pgpool seems not to be working
>> Sep 27 16:15:00 test01 pgpool[17854]: wd_lifecheck: lifecheck failed 3
>> times. pgpool seems not to be working
>> Sep 27 16:15:12 test01 pgpool[17854]: wd_lifecheck: lifecheck failed 3
>> times. pgpool seems not to be working
>> Sep 27 16:15:24 test01 pgpool[17854]: wd_lifecheck: lifecheck failed 3
>> times. pgpool seems not to be working
>> Sep 27 16:15:33 test01 pgpool[17847]: received smart shutdown request
>>
>> So this instance of pgpool start in backup mode and everything is
>> working fine (16:11). Then at 16:14 I shutdown the other pgpool
>> instance to test the failover, initially it works fine. This instance
>> activates the clustered IP and accepts connections. But then starting
>> at 16:14:48 the lifecheck fails for some strange reason and pgpool
>> blocks incoming connections. So I then shut it down.
>
> A small update, it seems that although I'm getting the lifecheck
> errors, pgpool is actually still working as it should. But it is still
> spamming the logs, so something is going wrong somewhere.
>
> I've enabled debug logging, but rsyslog seemed to have a issue with
> the amount of messages that produced. So I'll look into that.
>
> Regards,
> Tim
>
>


More information about the pgpool-general mailing list