pg_dump: could not format inet value

From: Tomas <tomas(dot)simonaitis(at)gmail(dot)com>
To: pgsql-general(at)postgresql(dot)org
Subject: pg_dump: could not format inet value
Date: 2007-12-05 09:04:20
Message-ID: 200712051104.20778.tomas.simonaitis@gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Hi,

I've stumbled on pg_dumpall problem:
--
pg_dump: ERROR: could not format inet value: Address family not supported by
protocol
pg_dump: SQL command to dump the contents of table "elog" failed: PQendcopy()
failed.
pg_dump: Error message from server: ERROR: could not format inet value:
Address family not supported by protocol
pg_dump: The command was: COPY public.elog (elog_id, elog_time, elog_src,
elog_userid, elog_type, elog_event, elog_trid) TO stdout;
pg_dumpall: pg_dump failed on database "logs", exiting
--
(elog_src column is only inet in elog table)

This problem happens on warm-standby server, which is periodically started to
dumpall and then goes back to PITR replication.
I cannot reliably reproduce the problem (but it happends every 30 dumps or
so).

Rows from elog table are never updated/deleted only inserted, so whatever
value might be problem it's there on next dump (which succeeds).

I've also inspected partial (from failed pg_dumpall) dump files - last
elog_src dumped and ones around it are normal IPv4 addresses.

postgresql-8.1 (8.1.9-0etch1) on both master and PITR slave.

postgres.log (start; pg_dumpall; stop)
--
2007-12-05 00:08:06 EET LOG: database system was interrupted at 2007-12-05
00:06:59 EET
2007-12-05 00:08:06 EET LOG: starting archive recovery
2007-12-05 00:08:06 EET LOG: restore_command
= "cp /var/lib/postgresql/8.1/backup/wals/%f %p"
cp: cannot stat `/var/lib/postgresql/8.1/backup/wals/00000001.history': No
such file or directory
2007-12-05 00:08:06 EET LOG: restored log
file "0000000100000019000000C9.0026A500.backup" from archive
cp: cannot stat
`/var/lib/postgresql/8.1/backup/wals/0000000100000019000000C9': No such file
or directory
2007-12-05 00:08:06 EET LOG: checkpoint record is at 19/C926A500
2007-12-05 00:08:06 EET LOG: redo record is at 19/C926A500; undo record is at
0/0; shutdown FALSE
2007-12-05 00:08:06 EET LOG: next transaction ID: 107741973; next OID: 44375
2007-12-05 00:08:06 EET LOG: next MultiXactId: 1; next MultiXactOffset: 0
2007-12-05 00:08:06 EET LOG: automatic recovery in progress
2007-12-05 00:08:06 EET LOG: redo starts at 19/C926A550
2007-12-05 00:08:06 EET LOG: incomplete startup packet
2007-12-05 00:08:06 EET LOG: record with zero length at 19/C9287558
2007-12-05 00:08:06 EET LOG: redo done at 19/C9287528
cp: cannot stat
`/var/lib/postgresql/8.1/backup/wals/0000000100000019000000C9': No such file
or directory
2007-12-05 00:08:06 EET LOG: archive recovery complete
2007-12-05 00:08:06 EET LOG: database system is ready
2007-12-05 00:08:06 EET FATAL: the database system is starting up
2007-12-05 00:08:07 EET LOG: transaction ID wrap limit is 1101205518, limited
by database "logs"
2007-12-05 00:08:40 EET ERROR: could not format inet value: Address family
not supported by protocol
2007-12-05 00:08:40 EET LOG: unexpected EOF on client connection
2007-12-05 00:10:15 EET LOG: incomplete startup packet
2007-12-05 00:10:15 EET LOG: received fast shutdown request
2007-12-05 00:10:15 EET LOG: shutting down
2007-12-05 00:10:15 EET LOG: database system is shut down
---

Browse pgsql-general by date

  From Date Subject
Next Message Andrus 2007-12-05 09:32:59 Re: Avoid huge perfomance loss on string concatenation
Previous Message Henrik Zagerholm 2007-12-05 08:57:56 Nested loop in simple query taking long time