Re: Re: Backup and Recovery - Mailing list pgsql-hackers

From Rod Taylor
Subject Re: Re: Backup and Recovery
Date
Msg-id 000b01c103f8$14634f70$2205010a@jester
Whole thread Raw
In response to Re: Backup and Recovery  (Matthew Kirkwood <matthew@hairy.beasts.org>)
List pgsql-hackers
> With stock PostgreSQL... how many committed transactions can one
lose
> on a simple system crash/reboot? With Oracle or Informix, the answer
> is zero. Is that true with PostgreSQL in fsync mode? If not, does it
> lose all in the log, or just those not yet written to the DB?

With WAL the theory is that it will not lose a committed transaction.
Bugs have plagged previous versions (7.1.2 looks clean) and it none
(Oracle, Informix, Postgres) can protect against coding errors in the
certain cases but from general power failure it's fine.

This assumes adequate hardware too.  Some harddrives claim to have
written when they haven't among other things, but Postgres itself
won't lose the information -- your hardware might :do that silently
though.)



pgsql-hackers by date:

Previous
From: ncm@zembu.com (Nathan Myers)
Date:
Subject: Re: Buffer access rules, and a probable bug
Next
From: Jan Wieck
Date:
Subject: Re: funny (cache (?)) bug in postgres (7.x tested)