Re: Rework the way multixact truncations work

On 2015-06-26 14:48:35 -0300, Alvaro Herrera wrote:
> Andres Freund wrote:
>
> > Rework the way multixact truncations work.
>
> I spent some time this morning reviewing this patch and had some
> comments that I relayed over IM to Andres.

Thanks for that!

> 2. We set PGXACT->delayChkpt while the truncation is executed. This
> seems reasonable, and there's a good reason for it, but all the other
> users of this facility only do small operations with this thing grabbed,
> while the multixact truncation could take a long time because a large
> number of files might be deleted. Maybe it's not a problem to have
> checkpoints be delayed by several seconds, or who knows maybe even a
> minute in a busy system. (We will have checkpointer sleeping in 10ms
> intervals until the truncation is complete).

I don't think this is a problem. Consider that we're doing all this in
the checkpointer today, blocking much more than just the actual xlog
insertion. That's a bigger problem, as we'll not do the paced writing
during that and such. The worst thatthis can cause is a bunch of sleeps,
that seems fairly harmless.

Greetings,

Andres Freund