Gaetano Mendola <mendola@bigfoot.com> writes:
> Tom Lane wrote:
>> "Christian Brosig" <Brosig@dataconsult.de> writes:
>>> Out of buffer error - Time to abort
>>
>> If this is 7.3.something, update to 7.3.4.
> I suspect that is an pgadmin III related problem,
> not a postgres problem.
No, it's definitely a backend failure. I think this CVS entry is
probably relevant:
2003-03-23 16:55 tgl
* src/backend/access/index/indexam.c (REL7_3_STABLE): Band-aid
patch for Shraibman's 'out of free buffers' bug: disable the
keys_are_unique optimization introduced in 7.3. A better fix will
appear in 7.4, but I think back-patching it is too risky for the
stable branch.
but if Christian is running something other than 7.3-7.3.2 then we need
another theory.
regards, tom lane