PostgreSQL Discussion - Had a problem with pg

Author

Had a problem with pg_clog

Peter Wiersig

2005-04-21, 8:24 pm

Hi List,

I had a database with a 8 Mb pg_dump file with 1.7 Gb in PG_DATA.

I dropped the database and tried to recreate the db with this
step:

psql template1 < data.sql

All tables and all data was restored and to help accessing the db
I tried to vacuum analyze it, yielding this error message:

vacuumdb -z miwabar
PANIC: open of /var/lib/pgsql/data/pg_clog/0000 failed:
Keine Berechtigung
server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
connection to server was lost
vacuumdb: vacuum miwabar failed

The panic message with -v gave
PANIC: read of clog file 0, offset 16384 failed: Erfolg

which lead me to the mailing-list archive and the following fix:

dd bs=32k count=1 </dev/zero >/var/lib/pgsql/data/pg_clog/0000
1+0 records in
1+0 records out

afterwards my vacuum works.

Had I done the right thing?

Should I have expected this error?

--
Peter

---------------------------(end of broadcast)---------------------------
TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql
.org

Tom Lane

2005-04-23, 8:23 pm

Peter Wiersig <peter@friesenpeter.de> writes:
> vacuumdb -z miwabar
> PANIC: open of /var/lib/pgsql/data/pg_clog/0000 failed:
> Keine Berechtigung
> server closed the connection unexpectedly

What PG version is this exactly? We have had some bugs of that
ilk in the past, but AFAIK they are all resolved in latest releases.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddressHere
" to majordomo@postgresql
.org)

Peter Wiersig

2005-04-24, 7:23 am

On Sat, Apr 23, 2005 at 08:19:31PM -0400, Tom Lane wrote:
> Peter Wiersig <peter@friesenpeter.de> writes:
>
> What PG version is this exactly?

It's 7.3.9 from SUSE rpms postgresql-7.3.9-3

--
Peter

---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings

Tom Lane

2005-04-24, 11:23 am

Peter Wiersig <peter@friesenpeter.de> writes:
> On Sat, Apr 23, 2005 at 08:19:31PM -0400, Tom Lane wrote:
[color=darkred]
> It's 7.3.9 from SUSE rpms postgresql-7.3.9-3

Hmm. 7.3.9 has all the known patches for hint-bit updates and premature
clog truncation, but maybe you have found a new bug. Can you repeat
this failure from a standing start --- that is, initdb, load your data
dump, vacuumdb, PANIC? If so I would very much appreciate a copy of
your data dump, if that's possible.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings

Peter Wiersig

2005-04-27, 1:23 pm

On Sun, Apr 24, 2005 at 11:52:11AM -0400, Tom Lane wrote:
> Peter Wiersig <peter@friesenpeter.de> writes:
>
>
> Hmm. 7.3.9 has all the known patches for hint-bit updates and
> premature clog truncation,

Good.

> Can you repeat this failure from a standing start --- that is,
> initdb, load your data dump, vacuumdb, PANIC?

No, I couldn't.

But I can with the old cluster.

And only if I create the new database with the sequence "psql
template1 <dump_with_create.sql; vacuumdb test".

Other combinations like "createdb test; psql test
<dump_with_create.sql; vacuumdb test" work.

I never tried to restore from a dump created by "pg_dump -C"
before, but this time I was in a hurry to free up more disk space.

--
Peter

---------------------------(end of broadcast)---------------------------
TIP 6: Have you searched our list archives?

http://archives.postgresql.org

Sponsored Links