On Thu, Apr 21, 2011 at 7:27 AM, Tomas Vondra <tv@fuzzy.cz> wrote:
> Dne 20.4.2011 22:11, Tomas Vondra napsal(a):
>> There's a very nice guide on how to do that
>>
>> http://blog.endpoint.com/2010/06/tracking-down-database-corruption-with.html
>>
>> It sure seems like the problem you have (invalid alloc request etc.).
>> The really annoying part is locating the block, as you have to scan
>> through the table (which sucks with such big table).
>>
>> And yes, if there's corruption, there might be more corrupted blocks.
>
> BTW, there's a setting 'zero_damaged_pages' that might help with this
>
> http://www.postgresql.org/docs/8.2/static/runtime-config-developer.html
>
> see this talk for more details how to use it
>
> http://www.casitconf.org/casitconf11/Tech_track_2_files/cascadia_postgres_rbernier.pdf
>
> Anyway don't play with this without the file backup, as this will zero
> the blocks.
>
> Tomas
Thanks Tomas. Very handy info.
FIRST: is there anyone on this list who offers PG admin support?
Please write to me directly.
Second, for the strace, which process should I use?
ps auxwww|grep ^postgres
postgres 4320 0.0 0.1 440192 10824 ? Ss 08:49 0:00
/usr/bin/postmaster -p 5432 -D /var/lib/pgsql/data
postgres 4355 0.0 0.0 11724 964 ? Ss 08:49 0:00
postgres: logger process
postgres 4365 0.0 0.0 440396 3268 ? Ss 08:49 0:00
postgres: writer process
postgres 4366 0.0 0.0 11860 1132 ? Ss 08:49 0:00
postgres: stats collector process
postgres 15795 0.0 0.0 7136 1440 pts/0 S 22:44 0:00 -bash
postgres 15900 0.0 0.0 7860 1956 pts/0 S+ 22:44 0:00 psql -h
localhost MYDOMAIN -E MYDOMAIN_MYDOMAIN
postgres 15901 0.0 0.0 441124 3072 ? Ss 22:44 0:00
postgres: MYDOMAIN_MYDOMAIN MYDOMAIN 127.0.0.1(34346) idle
Third, I have the backup in two ways:
1. I took a backup of the entire "/pgsql/data" folder. PG was shutdown
at the time.
2. I have a pg_dumpall file but it is missing one day's data (still
useful as last resort).
Will #1 have corrupt data in it?
