From: | Sam Mason <sam(at)samason(dot)me(dot)uk> |
---|---|
To: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Massively Parallel transactioning? |
Date: | 2010-08-19 12:41:57 |
Message-ID: | 20100819124157.GA2243@samason.me.uk |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Thu, Aug 19, 2010 at 05:40:21AM +0200, Adrian von Bidder wrote:
> On Thursday 19 August 2010 01.32:06 Benjamin Smith wrote:
> > This way we can be sure that either all the databases are in synch, or
> > that we need to rollback the program patch/update.
>
> I guess this might be more a hack than a solution: do the updates in batches
> and use 2pc: first connect to batches of databases, but instead of commit,
> you "prepare to commit". Prepared commits like this are persistent accross
> connections, so you can come back later and commit or rollback.
Rather than being a hack, 2PC sounds like it's needed for correctness;
how do you handle the case of only some databases receiving the COMMIT
command otherwise?
--
Sam http://samason.me.uk/
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2010-08-19 13:42:21 | Re: Win32 Backend Cash - pre-existing shared memory block is still in use |
Previous Message | darklow | 2010-08-19 11:48:20 | Re: FTS wildcard and custom ispell dictionary problem |