Database corruption

Aug 24 07:15:19 iprobe028 postgres[20034]: [3-1] ERROR:  could not read block 0 of relation 1663/16554/7463400: read only 0 of 8192 bytes

Aug 24 07:15:49 iprobe028 postgres[27663]: [3-1] ERROR:  could not read block 0 of relation 1663/16554/7463400: read only 0 of 8192 bytes

Aug 24 07:20:10 iprobe028 postgres[20231]: [4-1] ERROR:  could not open relation 1663/16554/7463400: No such file or directory

Aug 24 07:20:10 iprobe028 postgres[20231]: [5-1] ERROR:  current transaction is aborted, commands ignored until end of transaction block

Aug 24 07:20:18 iprobe028 postgres[20231]: [6-1] ERROR:  could not open relation 1663/16554/7463400: No such file or directory

Aug 24 07:20:51 iprobe028 postgres[25809]: [3-1] ERROR:  could not open relation 1663/16554/7463400: No such file or directory

Aug 24 07:20:52 iprobe028 postgres[20280]: [3-1] ERROR:  could not open relation 1663/16554/7463400: No such file or directory

Aug 24 07:21:52 iprobe028 postgres[20320]: [4-1] ERROR:  could not open relation 1663/16554/7463400: No such file or directory

Aug 24 07:22:52 iprobe028 postgres[20338]: [3-1] ERROR:  could not open relation 1663/16554/7463400: No such file or directory

Aug 24 07:23:52 iprobe028 postgres[20376]: [3-1] ERROR:  could not open relation 1663/16554/7463400: No such file or directory

We renamed and recreated the database that it was complaining about, and the error started up again with the new database within the hour.  We move the database to a different disk, with the same results.  The only way to fix it was to do a complete initdb on the box.

Tom: 

I did not see anything that might have been fixed in the 8.3 release notes that might be this issue.  I know this is an old version, but I have 2 copies of the data directory that should have the error if you (or whoever you delegate) would like to see them before I remove them.  I don't have an issue with sending the cluster data off list if it might be useful.

Thanks,

Woody