Message-ID: <1533281288.2479.5.camel@cybertec.at>
Subject: Re: PostgreSQL 9.3 - reindex problem
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: "For@ll" <forall@stalowka.info>, pgsql-admin@postgresql.org
Date: Fri, 03 Aug 2018 09:28:08 +0200
In-Reply-To: <pjuk85$u34$1@blaine.gmane.org>
References: <pjuk85$u34$1@blaine.gmane.org>
Organization: Cybertec
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

For@ll wrote:
> I have problem with postgresql-9.3 in our clients. They suddenly lost 
> power connection and all servers halt.. I repiar xfs filesystem on 
> CentOS7 postgresql partition. Now when I want run reindex database I 
> receive error:
> 
> reindexdb -d ex200
> NOTICE:  table "pg_catalog.pg_class" was reindexed
> reindexdb: reindexing of database "ex200" failed: ERROR:  invalid page 
> in block 101 of relation base/157934/12685
> 
> I want to try run dump and save database on plaintext but when I start 
> pg_dump it's hang...
> 
> pg_dump -a -v -d ex200 > ex200.sql
> pg_dump: last built-in OID is 16383
> pg_dump: reading extensions
> pg_dump: identifying extension members
> pg_dump: reading schemas
> pg_dump: reading user-defined tables
> pg_dump: reading user-defined functions
> 
> Any ideas what can I do.. Client told me they have a backup from 2 
> montgs ago....

That all looks like data corruption.

PostgreSQL should survive a crash just fine; it seems like your databases
are running with unsafe storage (that doesn't honor fsync).

If the backup from 2 months ago is a physical backup and you have WAL archives,
you can simply recover up to now.

If it is just a pg_dump, then you can either live with 2 months of data loss
or you hire professionaly that might be able to get some of your data back.

If you go for the second option, take a copy of the broken database
before doing anything else with it.

Yours,
Laurenz Albe
-- 
Cybertec | https://www.cybertec-postgresql.com