Re: Postgres 9.1: Adding rows to table causing too much latency in other queries

Jesper Krogh <jesper(at)krogh(dot)cc> writes:
> I have to say that I consistently have to turn "fastupdate" off for
> our heavily updated gin-indexes. The overall performance gain
> may be measurable, but its not intolerable without. The spikes seen
> from the applications, when cleanup happens. Either in the foreground
> or in the background are not tolerable. (multiple seconds).

Well, that's why there's a provision to turn it off: if response time
spikes are a bigger deal to you than overall performance, you probably
don't want bulk updates.

The theory is that you should be able to tune things so that the bulk
updates are done by autovacuum, but if you can't get that to work
sufficiently reliably, fastupdate=off is the best answer.

regards, tom lane