Title: Continuous archiving fails routinely with "invalid magic number" error
I am having problems getting continuous archiving to work reliably on my Postgres 8.3 system. My primary is running 8.3.1, has archiving turned on and writes logs to a NFS mount that lives on the backup server. The warm standby is running 8.3.3. Initially things work fine with the primary copying logs to the archive directory correctly and the warm standby in recovery mode plays them back okay. At some point the following appears in the syslog:
May 26 08:58:37 ods2-prod postgres[3827]: [1424-1] LOG: restored log file "00000001000000C2000000E3" from archive
May 26 08:58:51 ods2-prod postgres[3827]: [1425-1] LOG: restored log file "00000001000000C2000000E4" from archive
May 26 08:58:52 ods2-prod postgres[3827]: [1426-1] LOG: restored log file "00000001000000C2000000E5" from archive
May 26 08:58:55 ods2-prod postgres[3827]: [1427-1] LOG: invalid magic number 0000 in log file 194, segment 229, offset 10780672
May 26 08:58:55 ods2-prod postgres[3827]: [1428-1] LOG: redo done at C2/E5A45C80
May 26 08:58:55 ods2-prod postgres[3827]: [1429-1] LOG: last completed transaction was at log time 2009-05-26 08:58:21.804665+00
May 26 08:58:56 ods2-prod postgres[3827]: [1430-1] LOG: restored log file "00000001000000C2000000E5" from archive
May 26 08:59:11 ods2-prod postgres[3827]: [1431-1] LOG: selected new timeline ID: 2
May 26 08:59:26 ods2-prod postgres[3827]: [1432-1] LOG: could not create archive status file "pg_xlog/archive_status/00000002.history.ready": No such file or directory
May 26 08:59:26 ods2-prod postgres[3827]: [1433-1] LOG: archive recovery complete
I’ve had this occur three times since setting up the warm standby. The log file looks intact and I can’t see any errors in the primary’s syslog at the time the log is generated and copied. There are no NFS or space issues either on the backup server either. Can anyone suggest how to go about diagnosing this problem?
Bruce
