[pgpool-hackers: 3339] Re: [pgpool-committers: 5734] pgpool: Fix for duplicate failover request ...

[Tatsuo Ishii]

> On Mon, May 27, 2019 at 10:25 AM Tatsuo Ishii <ishii at sraoss.co.jp> wrote:
> 
>> > Hello, Usama and pgpool-hackers
>> >
>> > Thank you very much for writing and committing patches to solve the issue
>> > that I reported.
>> >
>> > What do you think about back-patching them into 3.7.x or older versions?
>> > Pgpool-II watchdog cluster can easily fall into the status that doesn't
>> > work without the patch.
>> >
>> > The quarantine feature hasn't been documented in detail up to now, so
>> > I think the behavior change of the first patch (*1) has no problem.
>> > My customer who has trouble related to this mail thread uses 3.7.x
>> version
>> > and strongly hopes such a pgpool improvement.
>> >
>> > *1
>> >
>> https://git.postgresql.org/gitweb/?p=pgpool2.git;a=commitdiff;h=3dd1cd3f1528
>>
>> We know that this patch breaks existing regression test. So I don't
>> think no user would notice the behavior change if the patch is
>> back ported.
>>
> 
> I sort of agree with Takatsuka San, that the continuous health checking on
> the quarantined backed nodes (first patch)
> is not the kind of change that should cause problems if we decide to
> backport it to older branches, In fact. it should be treated
> as a bug and back-patched to all affected branches. Also, the reason for
> the test case failure happening after that commit
> was not because of the behaviour or functionality change. rather it was due
> to the missing facility in health-check
> debugging/testing mechanism which was required by that test case after the
> change.

Alright. Then let's back port the firsr patch.

Best regards,
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese:http://www.sraoss.co.jp