Postgres 8.2 database recovery Could not create relation Invalid Argument

Started by Justin Alstonover 16 years ago2 messagesgeneral
Jump to latest
#1Justin Alston
jusalston@gmail.com

Novice here :). I have PostgreSQL 8.2 installed on a single board computer
running Windows XP Embedded on a Compact Flash drive - 2 databases with no
more than 2000 ro. After 10 power cycles spaced 6 mins apart, I noticed the
postgres.exe processes no longer running. I located log file (see below) and
it appears postgres attempts to recover but fails to do so. Any suggestions:

2009-07-28 04:31:01 LOG: database system was interrupted at 2009-07-28
04:27:54 GMT Daylight Time
2009-07-28 04:31:01 LOG: checkpoint record is at 0/7CBF58
2009-07-28 04:31:01 LOG: redo record is at 0/7CBF58; undo record is at 0/0;
shutdown TRUE
2009-07-28 04:31:01 LOG: next transaction ID: 0/12290; next OID: 17183
2009-07-28 04:31:01 LOG: next MultiXactId: 1; next MultiXactOffset: 0
2009-07-28 04:31:01 LOG: database system was not properly shut down;
automatic recovery in progress
2009-07-28 04:31:02 LOG: redo starts at 0/7CBFA8
2009-07-28 04:31:02 FATAL: could not create relation 1663/16403/16586:
Invalid argument
2009-07-28 04:31:02 CONTEXT: xlog redo update: rel 1663/16403/16586; tid
35/32; new 35/33
2009-07-28 04:31:02 LOG: startup process (PID 1484) exited with exit code 1
2009-07-28 04:31:02 LOG: aborting startup due to startup process failure
2009-07-28 04:31:02 LOG: logger shutting down

Thanks in advance for any insights.

Jus

#2Magnus Hagander
magnus@hagander.net
In reply to: Justin Alston (#1)
Re: Postgres 8.2 database recovery Could not create relation Invalid Argument

On Tuesday, July 28, 2009, Justin Alston <jusalston@gmail.com> wrote:

Novice here :). I have PostgreSQL 8.2 installed on a single board computer running Windows XP Embedded on a Compact Flash drive - 2 databases with no more than 2000 ro. After 10 power cycles spaced 6 mins apart, I noticed the postgres.exe processes no longer running. I located log file (see below) and it appears postgres attempts to recover but fails to do so. Any suggestions:

Any non standard config options? Particularly around fsync or or sync_method?

If not, I'd say it looks like a filesystem corruption issue. I assume
it's ntfs? How do you shut the machine down?

/Magnus

2009-07-28 04:31:01 LOG:  database system was interrupted at 2009-07-28 04:27:54 GMT Daylight Time
2009-07-28 04:31:01 LOG:  checkpoint record is at 0/7CBF58
2009-07-28 04:31:01 LOG:  redo record is at 0/7CBF58; undo record is at 0/0; shutdown TRUE
2009-07-28 04:31:01 LOG:  next transaction ID: 0/12290; next OID: 17183
2009-07-28 04:31:01 LOG:  next MultiXactId: 1; next MultiXactOffset: 0
2009-07-28 04:31:01 LOG:  database system was not properly shut down; automatic recovery in progress
2009-07-28 04:31:02 LOG:  redo starts at 0/7CBFA8
2009-07-28 04:31:02 FATAL:  could not create relation 1663/16403/16586: Invalid argument
2009-07-28 04:31:02 CONTEXT:  xlog redo update: rel 1663/16403/16586; tid 35/32; new 35/33
2009-07-28 04:31:02 LOG:  startup process (PID 1484) exited with exit code 1
2009-07-28 04:31:02 LOG:  aborting startup due to startup process failure
2009-07-28 04:31:02 LOG:  logger shutting down

Thanks in advance for any insights.

Jus

--
Magnus Hagander
Self: http://www.hagander.net/
Work: http://www.redpill-linpro.com/