Re: Recent buildfarm failures involving statement_timeout - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Recent buildfarm failures involving statement_timeout
Date
Msg-id 12227.1209392782@sss.pgh.pa.us
Whole thread Raw
In response to Re: Recent buildfarm failures involving statement_timeout  (Stefan Kaltenbrunner <stefan@kaltenbrunner.cc>)
List pgsql-hackers
Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> writes:
>> sysname  |      snapshot       |    branch   
>> ---------+---------------------+---------------
>> lionfish | 2007-04-19 09:30:27 | REL8_2_STABLE
>> lionfish | 2007-05-29 23:30:07 | REL8_1_STABLE
>> lionfish | 2007-09-22 23:30:07 | REL8_1_STABLE

> lionfish is(or rather was - powersupply got fried and I have no 
> replacement yet) one of the slowest machines(250MHZ mipsel,48MB RAM) on 
> the buildfarm (took about 5-6h for a complete run).

Unless someone has a better idea, I suggest we just raise the timeouts
used in prepared_xacts from 1 sec to 2 sec.  That should reduce the
recurrence rate to something vanishingly small if the problem is just
"it's too slow".  If it's that something's actually getting stuck, then
we'll continue to see failures and will know that closer investigation
is needed.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Andrew Sullivan
Date:
Subject: Re: we don't have a bugzilla
Next
From: Alvaro Herrera
Date:
Subject: Re: [PATCHES] Removing typename from A_Const (was: Empty arrays with ARRAY[])