Thread: catalog is missing 16 attribute(s) for relid 8202590
hello, i found this ERROR msg at our postgresql log file at one of my client.
what i did is i
1) kill -9 (postmaster process id).
2) remove the postmaster.pid file under /usr/local/pgsql/data/
3) restart the postmaster process.
then no more error msg in the postgresql log file.
problem come back few day later cause when i check the log file this morning, i saw the msg again.
the error msg will come out when we try to read record from the temporary table that created before.
so, i suspect the relid is refer to temporary table.(relation id), right?
what i did just restart the postmaster process should not be enough to solve the problem.. can someone give some advise on how to solve this problem?
Client Server Configuration
Postgresql 7.4.5 version.
Red Hat enterprise server 3.0 version
RAM 512MB
Dell PowerEdge 1800.
Regards
Beh
"Chun Yit\(Chronos\)" <ivanbeh@chronos.com.my> writes:
> what i did is i=20
> 1) kill -9 (postmaster process id).
> 2) remove the postmaster.pid file under /usr/local/pgsql/data/
> 3) restart the postmaster process.
The *first* thing you ought to do is learn not to do that. You
deliberately broke the safety interlocks that keep two separate sets of
backends from being active in a single database. If there's corrupted
data in there now, I think it's your own fault.
regards, tom lane
> i think i make a big misstake by using kill -9 the postmaster .so ,what > should i do when i facing this problem, can someone give advise, cause i > quite new to postgresql. > > how to check there is corrupted data in my database? > > it there anyway i can do to prevent this error msg come back > again? > > cause it just temporary table, can i just remove from database? Regards Beh ----- Original Message ----- From: "Tom Lane" <tgl@sss.pgh.pa.us> To: "Chun Yit(Chronos)" <ivanbeh@chronos.com.my> Cc: <pgsql-general@postgresql.org> Sent: Monday, May 08, 2006 10:43 PM Subject: Re: [GENERAL] catalog is missing 16 attribute(s) for relid 8202590 > "Chun Yit\(Chronos\)" <ivanbeh@chronos.com.my> writes: >> what i did is i=20 >> 1) kill -9 (postmaster process id). >> 2) remove the postmaster.pid file under /usr/local/pgsql/data/ >> 3) restart the postmaster process. > > The *first* thing you ought to do is learn not to do that. You > deliberately broke the safety interlocks that keep two separate sets of > backends from being active in a single database. If there's corrupted > data in there now, I think it's your own fault. > > regards, tom lane > > ---------------------------(end of broadcast)--------------------------- > TIP 6: explain analyze is your friend >