On 05/03/2014 04:29 PM, DrakoRod wrote:
> Hi everybody!
>
> I have a problem (really huge problem), I have one server of production, but
> yesterday in the night I saw this error:
>
> *ERROR: could not access status of transaction 2410303155
> DETAIL: Could not open file "pg_clog/08FA": No such file or directory*
>
> Solution:
> * dd if=/dev/zero/ of=data/pg_clog/08FA bs=256K count=1*
>
>
> So I ran this solutions, no problem so far, but then (after 1 or 2 hours
> approximately), the server crash, I think somebody did something but did not
> tell me, no one did nothing! cool!! ¬¬. I started the server and this began
> in recovery mode, he started. But after some time (without apparent
> pattern), the server came into the recovery mode again.
>
> After that, the server continuously entering recovery mode, like I said
> without apparent pattern, between 3, 5 or 10 minutes run normally but then
> enter in the recovery mode again. I restart the server (began in recovery
> mode again), started, but after sometime he enter in recovery mode again.
>
> Try to recover the server with the PITR and nothing. The server version is
> 9.0.x in a Linux SUSE. The database size is the 336 GB.
>
> Please give me any help to recover the server!
In addition to what David said about the log information, do you have
any program running against the pg_clog/ directory?
Something that might be moving the files out of it?
>
> Thanks!
>
>
--
Adrian Klaver
adrian.klaver@aklaver.com
