Re: Re: Better way of dealing with pgstat wait timeout during buildfarm runs?

On 2015-01-19 11:16:09 -0500, Tom Lane wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> >> I'd be fine with changing the warning to LOG level rather than
> >> suppressing it entirely for the specific case of pgstat_vacuum_stat;
> >> but I do want to distinguish that case, wherein it's fair to conclude
> >> that obsolete stats aren't too worrisome, from other cases where no
> >> such conclusion is justified.
>
> > But I can live with this compromise.
>
> Is that OK with everybody? Going once, going twice ...

I can live with the compromise as well, but I'd rather change it to
always be of LOG priority. LOG is more likely to end up in the log and
that's where it's actually likely to be noticed. In most of the cases
WARNINGs going to the client won't be noticed as this error is much more
likely on servers with a high load caused by programs than during
interactive use.

> > Sure, but nobody who is not a developer is going to care about that.
> > A typical user who sees "pgstat wait timeout", or doesn't, isn't going
> > to be able to make anything at all out of that.
>
> Possibly we need to improve the wording of that error message then.
> When it was written, we really assumed that it was a can't-happen case
> and so didn't spend much effort on it. Perhaps it should become a
> translatable ereport phrased like "WARNING: using stale statistics
> instead of current ones because stats collector is not responding".

Yes, that seems like a good message improvement.

It's not like making it LOG makes the message invisible...

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services