[pgpool-general: 4271] Re: Node Recovery

[Steven Livingstone]

So, no longer did I write this than i found a command
called pcp_attach_node that seems to get it all working. Typical!

However, given the complexity of some of the other solutions I am wondering
whether this isn't actually doing what I think it is ... specifically, with
respect to pcp_recovery_node.

steven

On Wed, Dec 30, 2015 at 9:16 PM, Steven Livingstone <steven at livz.org> wrote:

> Hello - this is my first real step into the world of Postgres and Pgpool.
>
> I have managed to get a simple two node Postgres master/slave cluster
> working, using Postgres Streaming Replication. Works well.
>
> I have then added Pgpool (one instance for now, tho looking at HA), added
> the master/slave above as backends and that works fine. Then I turn on load
> balancing, set master_slave_mode=true and sure enough, i can bring down one
> and the other takes over.
>
> However, when I bring the slave down, even when reattached it stays at
> status=3 which i know is node down.
>
> Now, I have researched and threads around this go back a number of years
> and much of what I tried didn't quite work ... i suspect it applies when I
> use PGpool streaming replication rather than the SR now part of Postgres.
>
> Does anyone have a fairly straightforward guide/suggestion on how to use
> the Postgres SR in Master/Slave config, along with PGpool when I can add
> nodes back (or rather make them available) into the pgpool?
>
> Been a fairly steep learning curve but hoping to sort it by next year ;-)
>
> thanks,
> steven
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20151230/936d81c1/attachment.html>