Re: pg_stop_backup(wait_for_archive := true) on standby server

On Thu, Jun 29, 2017 at 10:30 PM, Magnus Hagander <magnus(at)hagander(dot)net> wrote:
> On Thu, Jun 22, 2017 at 6:22 PM, Masahiko Sawada <sawada(dot)mshk(at)gmail(dot)com>
> wrote:
>>
>> On Thu, Jun 22, 2017 at 10:36 PM, Magnus Hagander <magnus(at)hagander(dot)net>
>> wrote:
>> >
>> >
>> > On Thu, Jun 22, 2017 at 10:12 AM, Masahiko Sawada
>> > <sawada(dot)mshk(at)gmail(dot)com>
>> > wrote:
>> >>
>> >> Hi,
>> >>
>> >> Since an optional second argument wait_for_archive of pg_stop_backup
>> >> has been introduced in PostgreSQL 10 we can choose whether wait for
>> >> archiving. But my colleagues found that we can do pg_stop_backup with
>> >> wait_for_archive = true on the standby server but it actually doesn't
>> >> wait for WAL archiving. Because this behavior is not documented and we
>> >> cannot find out it without reading source code it will confuse the
>> >> user.
>> >>
>> >> I think we can raise an error when pg_stop_backup with
>> >> wait_for_archive = true is executed on the standby. Attached patch
>> >> change it so that.
>> >
>> >
>> > Wouldn't it be better to make it *work*? If you have
>> > archive_mode=always, it
>> > makes sense to want to wait on the standby as well, does it not?
>> >
>>
>> Yes, ideally it will be better to make it wait for WAL archiving on
>> standby server when archive_mode=always. But I think it would be for
>> PG11 item, and this item is for PG10.
>>
>
> I'm not sure. I think this can be considered a bug in the implementation for
> 10, and as such is "open for fixing". However, it's not a very critical bug
> so I doubt it should be a release blocker, but if someone wants to work on a
> fix I think we should commit it.
>

I agree with you. I'd like to hear opinions from other hackers as well.

Regards,

--
Masahiko Sawada
NIPPON TELEGRAPH AND TELEPHONE CORPORATION
NTT Open Source Software Center