Re: [Linux-HA] PostGreSQL, Warm standby and STONITH

2009-04-17 Thread David Dumortier
Hi all, I forgot the list, sorry for cross-posting Serge. Serge Dubrouski a écrit : The problem is that I don't really understand what we are talking about here. Are you trying to build master/slave setup using SLONY or some other replication mechanism? If yes, then pgsql RA doesn't support

Re: [Linux-HA] PostGreSQL, Warm standby and STONITH

2009-04-16 Thread David Dumortier
Hi all, Dejan Muhamedagic a écrit : Hi, There is a RA for postgres (pgsql). I guess you should try to use that. I will have to modify it. The Warm standy of PG work with two servers running : 1 a master pushing data to slave 2 a slave taking data from master 3 when master is down a flag have

Re: [Linux-HA] PostGreSQL, Warm standby and STONITH

2009-04-16 Thread Dejan Muhamedagic
Hi, On Thu, Apr 16, 2009 at 11:57:28AM +0200, David Dumortier wrote: Hi all, Dejan Muhamedagic a ?crit : Hi, There is a RA for postgres (pgsql). I guess you should try to use that. I will have to modify it. The Warm standy of PG work with two servers running : 1 a master pushing

Re: [Linux-HA] PostGreSQL, Warm standby and STONITH

2009-04-16 Thread Serge Dubrouski
Don't modify it, write a new one. On Thu, Apr 16, 2009 at 3:57 AM, David Dumortier david.dumort...@linagora.com wrote: Hi all, Dejan Muhamedagic a écrit : Hi, There is a RA for postgres (pgsql). I guess you should try to use that. I will have to modify it. The Warm standy of PG work with

Re: [Linux-HA] PostGreSQL, Warm standby and STONITH

2009-04-15 Thread Dejan Muhamedagic
Hi, On Tue, Apr 14, 2009 at 05:03:55PM +0200, David Dumortier wrote: Hi the list, I have to setup an heartbeat on a 2 computers-cluster with PostGreSQL. I setuped a fail-over on PG with a warm-stanby : a file flag tell the slave to take the master state. There is a RA for postgres (pgsql).

[Linux-HA] PostGreSQL, Warm standby and STONITH

2009-04-14 Thread David Dumortier
Hi the list, I have to setup an heartbeat on a 2 computers-cluster with PostGreSQL. I setuped a fail-over on PG with a warm-stanby : a file flag tell the slave to take the master state. I'm new on high-availability. So I browse documentation, list and web but I've some questions. It seems