[Pgpool-general] Replication Flawed?
Andrew Sullivan
ajs at crankycanuck.ca
Wed May 17 13:55:25 UTC 2006
On Wed, May 17, 2006 at 03:02:25PM +0200, Robert Ayres wrote:
> I have run the same tests (and passed) using a JDBC based
> replication system, Sequoia (sequoia.continuent.org), which ensures
> replication consistency by ordering statements from all
> transactions based on ids assigned to them AFAIK.
Sequoia can't guarantee replication either. Try doing SELECT now()
on two machines, down to the millisecond. I bet you get very
slightly different answers. Or better, select oid from pg_class
where tname = [something].
Replication at the statement level is always prone to this sort of
problem. That isn't to say it's a bad thing: you just need to
evaluate the failure conditions against your application.
A
--
Andrew Sullivan | ajs at crankycanuck.ca
Unfortunately reformatting the Internet is a little more painful
than reformatting your hard drive when it gets out of whack.
--Scott Morris
More information about the Pgpool-general
mailing list