Hi,
I recently had set up WAL archiving in a testing environment and all went
well. As a test I let the disk fill up until archiving no longer was
possible.
Then a co-worker noticed this disk filling up and removed some WALL
segments, that were not yet archived. Now the archiving proces hangs.
Messages in logfile:
2007-12-04 15:20:50 MET LOG: archive command "cp
"pg_xlog/00000001000000010000003A"
/u002/postgresql/admin/mdbp/arch/"archive_wal_00000001000000010000003A""
failed: return code 256
2007-12-04 15:20:51 MET DEBUG: executing archive command "cp
"pg_xlog/00000001000000010000003A"
/u002/postgresql/admin/mdbp/arch/"archive_wal_00000001000000010000003A""
cp: cannot access pg_xlog/00000001000000010000003A: No such file or
directory
2007-12-04 15:20:51 MET LOG: archive command "cp
"pg_xlog/00000001000000010000003A"
/u002/postgresql/admin/mdbp/arch/"archive_wal_00000001000000010000003A""
failed: return code 256
2007-12-04 15:20:51 MET WARNING: transaction log file
"00000001000000010000003A" could not be archived: too many failures
How can I resolve this issue? I only have WALL segments from
000000010000000100000080 so I am missing all segments from
00000001000000010000003A upto 000000010000000100000080.
--
View this message in context:
http://www.nabble.com/WAL-archiving-hangs-due-to-missing-WAL-segment-tf4943617.html#a14152262
Sent from the PostgreSQL - admin mailing list archive at Nabble.com.