Alvaro Herrera wrote:
> Alvaro Herrera wrote:
>
>> Setting the cost_delay sounds a reasonable thing to do anyway, and in
>> fact I already proposed it and nobody objected (AFAIR). Now we only
>> have to agree on a reasonable value.
>
> Also note this message:
>
> Date: Sat, 27 Jan 2007 21:51:40 -0500
> Message-ID: <27665.1169952700@sss.pgh.pa.us>
> From: Tom Lane <tgl@sss.pgh.pa.us>
> Subject: Re: [GENERAL] Stats collector frozen?
>
> [...]
>> If this theory is correct, then we can improve the reliability of the
>> stats test a good deal if we put a sleep() at the *start* of the test,
>> to let any old backends get out of the way. It seems worth a try
>> anyway. I'll add this to HEAD and if the stats failure noise seems to
>> go down, we can back-port it.
>
> which was followed by this commit
>
> revision 1.6
> date: 2007-01-28 00:02:31 -0300; author: tgl; state: Exp; lines: +4 -0;
> Add a delay at the start of the stats test, to let any prior stats
> activity quiesce. Possibly this will fix the large increase in
> non-reproducible stats test failures we've noted since turning on
> stats_row_level by default.
>
>
> Apparently it wasn't enough to completely eliminate the problems. Did
> it reduce them? I haven't been watching the buildfarm closely enough to
> know for sure.
at least for my members it seems it did not have any effect at all.
I actually think I got more failures in the period afterwards but the
failures are too sporadic to quantify that(and in some way also depends
on the number of commits done which directly influence the number of
builds/tests).
Stefan
