Re: Reg: PostgreSQL Server "base/data" recovery

Thank you Andrew for your reply.

I see, As there are many reasons exists for the data corruption i couldn't
figure out it exactly.
Unfortunately there is no backup for this system and is not possible to
restore from the backup.

I have been using the NFS storage since 2013 and didn't experience this
issue before, now i am afraid.
We haven't set to run the fysnc() in our environment.

If i set "fysnc to on" now, did it make any impact to current flow?

Also i will considering the suggestion to upgrade the PostgreSQL.

Regards,
Prajilal

On Thu, Mar 19, 2015 at 7:39 PM, Andrew Sullivan <ajs(at)crankycanuck(dot)ca>
wrote:

> On Thu, Mar 19, 2015 at 07:02:28PM +0900, Prajilal KP wrote:
> >
> > When i see check the this file, the file itself exists but the size is
> "0"
> > byte.
>
> That suggests you have data corruption, and that you need to restore from
> backup.
>
> > The server is writing the whole log in to the mounted network storage,
> NFS.
>
> There are reasons that people get nervous about databases on NFS. Are
> you ensuring that Postgres fsync() calls (like when COMMIT happens)
> are not being handled asynchronously?
>
> Also, a trivial scan of the release notes in the 9.0.x series shows a
> number of data corruption fixes since 9.0.4. You should always try to
> stay on the latest minor release of your version of Postgres.
>
> Best regards,
>
> A
>
> --
> Andrew Sullivan
> ajs(at)crankycanuck(dot)ca
>
>
> --
> Sent via pgsql-general mailing list (pgsql-general(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-general
>