Re: Information about XLogWrite - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Information about XLogWrite
Date
Msg-id 20376.1014152769@sss.pgh.pa.us
Whole thread Raw
In response to Information about XLogWrite  (Aniket Kulkarni <aniket_kulkarni@persistent.co.in>)
List pgsql-hackers
Aniket Kulkarni <aniket_kulkarni@persistent.co.in> writes:
> Recently I got this error while the database was starting up. :
> FATAL 2: XLogWrite: writerequest is past end of log
> And the database did not startup, and finally I had to do a
> pg_resetxlog. I tried going through the code which gave
> this error but I was unable to comprehend much.

> Could anyone tell me how exactly this error is caused.

XLogWrite?  Not XLogFlush?  That'd be a new one on me.  Too bad you
flushed the broken xlog, I would have liked to look at it.

There is a known failure path (believed fixed in 7.2) whereby garbage
data in pg_log pages can lead to XLogFlush startup failures.
        regards, tom lane


pgsql-hackers by date:

Previous
From: nconway@klamath.dyndns.org (Neil Conway)
Date:
Subject: Re: questions on ALTER TABLE ... OWNER
Next
From: Stephan Szabo
Date:
Subject: Re: pg_dump: use ALTER TABLE for PKs