[pgpool-hackers: 2059] Re: Pgpool-II should not perform ping test after bringing down the VIP

[Tatsuo Ishii]

> On Thu, Feb 23, 2017 at 5:45 AM, Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> 
>> Usama,
>>
>> I thought de-escalation and escalation happen sequentially, rather
>> than concurrently, but it seems I was wrong. If they happen in
>> sequential, you thought seems to be correct.

I meant "If they happen in parallel, you thought seems to be correct."

> The new leader election and escalation start as soon as the old master
> informs the cluster that it is going down, So escalation and de-escalation
> can happen in parallel and it is necessary to ensure minimum down time at
> the time of the master node switch.
> 
> I will make the necessary changes.

Please go ahead.

> Regards
> Muhammad Usama
> 
>>
>> Best regards,
>> --
>> Tatsuo Ishii
>> SRA OSS, Inc. Japan
>> English: http://www.sraoss.co.jp/index_en.php
>> Japanese:http://www.sraoss.co.jp
>>
>> > Hi Ishii-San
>> >
>> > Any thoughts on this one?
>> >
>> > Kind regards
>> > Muhammad Usama
>> >
>> > On Thu, Feb 16, 2017 at 2:05 PM, Muhammad Usama <m.usama at gmail.com>
>> wrote:
>> >
>> >> Hi
>> >>
>> >> At the time de-escalation from the master watchdog node, Pgpool-II
>> should
>> >> not perform the ping test to verify if the VIP is successfully brought
>> down
>> >> or not.
>> >>
>> >> The reason is, if the newly elected master watchdog node acquires the
>> VIP
>> >> while the resigning Pgpool-II node is still performing the de-escalation
>> >> steps or doing the ping test on VIP. The resigning Pgpool-II will
>> >> falsely assume that it has failed to bring down the delegate IP, after
>> >> getting the positive ping result from the new master node that has
>> already
>> >> acquired that VIP.
>> >>
>> >> Comments and suggestions are welcome.
>> >>
>> >> Kind regards
>> >> Muhammad Usama
>> >>
>>