[pgpool-general: 359] Re: Multiple pgpool servers failover

[Matt Wise]

Lou,
  I'm afraid that I want to have the same design as you, but it simply doesnt exist right now in PGPool. Right now the behavior is EITHER:

a) failover is allowed, and the instance any postgres server goes down the appropriate failover scripts are called
b) failover is DISALLOWED, and the instant any postgres server goes down, your entire pgpool server basically freezes all connections until the servers come back up.

  There's no middle ground... I'm actually talking to one of my engineers in house and we're thinking about patching PGPool so that we can have a behavior where:

If DISALLOW_TO_FAILOVER is set and ..
  a) one of the slaves fails: that slave is taken out of rotation until its back up
  b) the master fails: all connections hang until the master is back, OR a new master is detected. pgpool goes into a loop looking for new masters.

—Matt

On Apr 19, 2012, at 6:27 PM, Lou Kamenov wrote:

> Hey there,
> 
> I have been looking at a setup which includes couple of pgpool servers
> and about four PostgreSQL nodes in a 1 master, 3 slaves streaming
> replication configuration.
> 
> From what I've seen PgPool performs the fallback to backend1 the
> moment the master disappears, but in case of two pgpool nodes which
> point to the same set of servers, what would the behavior be?
> 
> In effect, the first pgpool that detects the failing master will do
> the switch and trigger the slave into master mode, what would the
> other pgpool do, attempt to trigger backend1 into master mode as well?
> will it detect that the master was already triggered, etc.
> 
> I will be trying this in practice but would be great to hear other
> peoples experience.
> 
> Thanks in advance!
> Lou
> _______________________________________________
> pgpool-general mailing list
> pgpool-general at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-general

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20120419/5cc62d7e/attachment.html>