On 31.05.2012 08:06, Erik Rijkers wrote:
> On Thu, May 31, 2012 03:30, Robert Haas wrote:
>> On Wed, May 30, 2012 at 6:00 PM, Erik Rijkers<er@xs4all.nl> wrote:
>>> directory
>>> 2012-05-30 23:40:57.909 CEST 3909 CONTEXT: writing block 5152 of relation base/21268/26569
>>> xlog redo multi-insert (init): rel 1663/21268/26581; blk 3852; 35 tuples
>>> TRAP: FailedAssertion("!(PrivateRefCount[i] == 0)", File: "bufmgr.c", Line: 1741)
>>> 2012-05-30 23:40:58.006 CEST 5331 FATAL: could not open file "base/21268/26569": No such file
>>> or
>>> directory
>>> 2012-05-30 23:40:58.006 CEST 5331 CONTEXT: writing block 5153 of relation base/21268/26569
>>> 2012-05-30 23:40:59.661 CEST 3908 LOG: startup process (PID 3909) was terminated by signal 6:
>>> Aborted
>>> 2012-05-30 23:40:59.661 CEST 3908 LOG: terminating any other active server processes
>>
>> Hmm. I set up what I believe to be the same test case you were using,
>> and it's not crashing for me. In fact, with the latest code, I
>> haven't been able to produce any error at all. When I reverted my
>> last commit, I managed to get this:
>>
>> ERROR: could not open relation with OID 18229
>> STATEMENT: select current_setting('port') port, count(*) from t
>>
>> ...but just once, and with no other error messages. So I'm either
>> missing a step somewhere, or something about your system just happens
>> to tickle this moreso than on mine.
>
> In my test, I run the bash code (the bits that I posted earlier) in a while loop so that the table
> is CREATEd, COPYied into, and DROPped every few seconds -- perhaps that wasn't clear. That loop
> is necessary; a few iterations are often successful.
>
> I can test it today on a few other systems to see if it is reproducible.
I could no longer reproduce it after Robert's fix, the test case has
been running for about an hour now. Please triple-check that you have it
applied in the standby :-).
-- Heikki Linnakangas EnterpriseDB http://www.enterprisedb.com