Re: Additional logging for VACUUM and ANALYZE

On 10/5/17, 12:29 AM, "Michael Paquier" <michael(dot)paquier(at)gmail(dot)com> wrote:
> On Thu, Oct 5, 2017 at 1:23 AM, Bossart, Nathan <bossartn(at)amazon(dot)com> wrote:
>> Presently, there are a few edge cases in vacuum_rel() and analyze_rel() that I
>> believe do not have sufficient logging. This was discussed a bit in the
>> vacuum-multiple-relations thread [0], but it was ultimately decided that any
>> logging changes should be proposed separately.
>
> I think that I agree with that, especially now with VACUUM allowing
> multiple relations. The discussion then would be how much logging we
> want. WARNING looks adapted per the discussions we had on the other
> thread as manual VACUUMs can now involve much more relations, even
> with partitioned tables. More opinions would be welcome.

Thanks for taking a look.

>> and a test that exercises a bit of this functionality.
>
> My take on those test would be to not include them. This is a lot just
> to test two logging lines where the relation has been dropped.

Yeah, I'm not terribly concerned about those tests.

>> If this change were to be considered for back-patching, we would likely want to
>> also apply Michael's RangeVar fix for partitioned tables to 10 [1]. Without
>> this change, log messages for unspecified partitions will be emitted with the
>> parent's RangeVar information.
>
> Well, that's assuming that we begin logging some information for
> manual VACUUMs using the specified RangeVar, something that does not
> happen at the top of upstream REL_10_STABLE, but can happen if we were
> to include the patch you are proposing on this thread for
> REL_10_STABLE. But the latter is not going to happen. Or did you patch
> your version of v10 to do so in your stuff? For v10 the ship has
> already sailed, so I think that it would be better to just let it go,
> and rely on v11 which has added all the facility we wanted.

I agree. I didn't mean to suggest that it should be back-patched, just
that v10 has a small quirk that needs to be handled if others feel
differently.

Nathan