Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> writes:
> Tom Lane wrote:
>> which makes it seem highly probable that this recently committed patch
>> to convert the MIPS out-of-line spinlock code into inline assembler
>> isn't right:
>> http://archives.postgresql.org/pgsql-committers/2005-08/msg00319.php
> As the owner of said machine I was about to report the problem - but on
> a subsequent run of the buildfarm-script(to get access to the compiled
> source for further debugging and testing) it completed without an error.
Unfortunately, that just says that the failure is intermittent, which is
pretty much exactly what you'd expect from not-quite-right locking code.
Try running the parallel regression tests several times in a row. (The
serial test mode is unlikely to exhibit any problem.) I tend to do this
by installing the test postmaster and then doing "make
installcheck-parallel" over and over, rather than repeating "make
check"; repeatedly building a test installation is just wasting time
for this sort of thing.
regards, tom lane
