[pgpool-general: 1931] Auto recovery of nodes and dealing with OS updates

[Nathan Brennan]

Hi,

I was just wonder how people do there OS updates as it seems like it could
be quiet a time consuming manual process.

I have a basic master/salve replication with auto failover and online
recovery which is great for dealing with issues that shouldn't happen often
eg server dying, but it seems to add some major over head into doing basic
things like OS updates.

For example if I was to do OS updates (security patches) on all the
servers( only 2 )  I would have to do the following:
1) stop postgres on server 1 assume this was primary and pgpool does
auto-failover to the standby
2) apply updates
3) start postgres up
4) pgpool now sees the node but it requires online recovery to turn it into
a slave and catchup on replication
5) now we stop postgres server 1 with does auto-failover back to server 1
6) apply updates
7) start postgres back up
8) pgpool now has to do online recovery again.

preferably during step 4 I would like to automatically promote the server
slave and catchup by reading the wal segments this is assume the node has
been down for less then x minutes, this would allow reliance on the
failover script for when I did updates which could be completely automated.

Is there some way of doing this eg pgpool see the server is back and rans a
script or does someone have a better solution :).

Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.sraoss.jp/pipermail/pgpool-general/attachments/20130724/53dfa7cd/attachment.html>