Обсуждение: Recovery error

Hi,

Accidentally I've ran out of disk space during intensive database write
operations. At the begining the backend went to recovery mode a few times,
and continued the work, but finally it has terribly died ( Segfault at backend/utils/mmgr/mcxt.c 222 )

After this event, the recovery has stoped  this way:

[version: postgres 7.1 beta6
OK, I know this is very old... I do the upgrade soon, but I am
interested in what this error message means, could this happen with the
newser versions and so on.....]


pgdevel@road:/home/postgreSQL/postgres7.1$ postgres -O -P -Ddata performance
DEBUG:  database system shutdown was interrupted at 2001-07-31 16:05:04 CEST
DEBUG:  CheckPoint record at (1, 2941129720)
DEBUG:  Redo record at (1, 2936012956); Undo record at (1, 2936012808); Shutdown FALSE
DEBUG:  NextTransactionId: 5699520; NextOid: 17895453
DEBUG:  database system was not properly shut down; automatic recovery in progress...
DEBUG:  redo starts at (1, 2936012956)
DEBUG:  open(logfile 1 seg 179) failed: No such file or directory
DEBUG:  redo done at (1, 3003121540)
FATAL 2:  XLogWrite: write request is past end of log
pgdevel@road:/home/postgreSQL/postgres7.1$


Attila

Meszaros Attila <tilla@draconis.elte.hu> writes:
> Accidentally I've ran out of disk space during intensive database write
> operations. At the begining the backend went to recovery mode a few times,
> and continued the work, but finally it has terribly died 
>     ( Segfault at backend/utils/mmgr/mcxt.c 222 )

A backtrace from that segfault might be interesting, if you can provide
one.

> [version: postgres 7.1 beta6
> OK, I know this is very old... I do the upgrade soon,

I suggest "now", not "soon" :-(.  The betas had some bugs in logfile
recovery, and I think you've been bit by one.  This looks familiar:

> FATAL 2:  XLogWrite: write request is past end of log

You might be able to recover your data by updating to 7.1.2 and running
contrib/pg_resetxlog before you start the new postmaster.
        regards, tom lane