Re: More along the lines of "WEIRD CRASH?!?!"?

Tom Lane [mailto:tgl(at)sss(dot)pgh(dot)pa(dot)us] writes:
Murthy Kambhampaty <murthy(dot)kambhampaty(at)goeci(dot)com> writes:
>> Mar 18 12:33:45 c257lx postgres[25277]: [7] PANIC: open of
>> /home/db/pgsql/data/pg_clog/0406 failed: No such file o
>> r directory
>
>> This always happens several passes into a loop (second or third with
default
>> shared memory settings, much later with 4096 shared buffers and 32
>> checkpoint segments).
>
>Is it always complaining of the same pg_clog file name? Are you getting
>rid of the corruption between tries (eg with initdb), or is it just that
>you keep hitting the same pre-existing corrupted data?

I am not running initdb between tries to get rid of the corruption (I am
dropping the tables that are created). However, the filename is not always
the same (the repeated ones are almost certainly cases where I forgot to
drop the tables between iterations):
# less /var/log/postgresql | grep clog
Mar 17 13:42:11 c257lx postgres[30988]: [7] PANIC: open of
/home/db/pgsql/data/pg_clog/0001 failed: No such file or directory
Mar 17 13:47:02 c257lx postgres[32111]: [10] PANIC: open of
/home/db/pgsql/data/pg_clog/0001 failed: No such file or directory
Mar 17 20:33:58 c257lx postgres[6483]: [7] PANIC: open of
/home/db/pgsql/data/pg_clog/0B02 failed: No such file or directory
Mar 17 23:52:12 c257lx postgres[8161]: [10] PANIC: open of
/home/db/pgsql/data/pg_clog/0B02 failed: No such file or directory
Mar 18 04:03:06 c257lx postgres[8628]: [13] PANIC: read of clog file 0,
offset 147456 failed: Success
Mar 18 09:25:28 c257lx postgres[10341]: [15] PANIC: read of clog file 0,
offset 196608 failed: Success
Mar 18 12:33:45 c257lx postgres[25277]: [7] PANIC: open of
/home/db/pgsql/data/pg_clog/0406 failed: No such file or directory
Mar 18 13:12:56 c257lx postgres[25398]: [69] PANIC: open of
/home/db/pgsql/data/pg_clog/0730 failed: No such file or directory

I can test with an initdb between iterations, if that would help. (I am in
the process of setting up 7.2.4.)

Thanks,
Murthy