Thread: Help! xlog flush request is not satisfied

Hi.
I just logged into a system and found this in the log when trying to
start postgres.

LOG:  database system shutdown was interrupted; last known up at
2009-12-07 06:27:33 CET
LOG:  database system was not properly shut down; automatic recovery in progress
LOG:  redo starts at 15/B320AF68
LOG:  record with zero length at 15/B348B9B0
LOG:  redo done at 15/B348B980
LOG:  last completed transaction was at log time 2009-12-07 05:54:19.655173+01
FATAL:  xlog flush request 15/B37CA000 is not satisfied --- flushed
only to 15/B348B9B0
CONTEXT:  writing block 0 of relation base/16386/119806_vm
LOG:  startup process (PID 27338) exited with exit code 1
LOG:  aborting startup due to startup process failure

And now I'm unsure what to do. (running postgresql 8.4.1)
Any idea on what is wrong here, any suggestions on what I should do?
I've found out that there is a command pg_resetxlog that some people
have run in these circumstances, but I'd rather not run anthing before
someone tells me what is going on.

There is a backup (pg_dumpall)  that is about 12 hours old that I
think I could use if that is the best option.

On 7/12/2009 3:00 PM, A B wrote:
> Hi.
> I just logged into a system and found this in the log when trying to
> start postgres.

Possibly silly question: is there plenty of free disk space on the file
system where PostgreSQL's data is stored?

--
Craig Ringer

Craig Ringer <craig@postnewspapers.com.au> writes:
> On 7/12/2009 3:00 PM, A B wrote:
>> I just logged into a system and found this in the log when trying to
>> start postgres.

> Possibly silly question: is there plenty of free disk space on the file
> system where PostgreSQL's data is stored?

That's not what that message is about.  What it's unhappy about is that
it found a page with an LSN that's past the logical end of WAL.  This
either means that page is corrupt, or there's something wrong with WAL.

            regards, tom lane