|
Hi Peter!
Thanks for your reply!
----- Original Message -----
From: "Peter Mueller" <pmueller@xxxxxxxxxxxx>
To: <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, December 26, 2002 9:26 PM
Subject: RE: Server Application with Primary Server & Backup Server
> Gregor,
>
> > 1. How can I trigger a failover between the primary server
> > and the backup
> > server?
> > Must I turn of heartbeat with /etc/init.d/heartbeat stop or
> > is ther another
> > availability to do this? (Could it be done with the Heartbeat API?)
>
> I asked the same question a while back and the response was to just
> stop/start the service.
>
> Not sure about the API, but hey that sounds like a reasonable thing for an
> API to do ;). Perhaps the heartbeat mailing list might provide a clue:
> http://marc.theaimsgroup.com/?l=linux-ha&r=1&w=2 <-- archive
> "
I hadn't read anything about the heartbeat API (only a interface
description, but nothing else)
> send a message to linux-ha-subscribe@xxxxxx to subscribe to the list.
> send a message to linux-ha-unsubscribe@xxxxxx to unsubscribe
> "
>
> > 2. If the primary server is back again, a handover from the
> > backup server is
> > triggered. Is it possible to prevent this behaviour? Or are there some
> > reasons for doing this?
>
> This is the nice_failback setting in ha.cf. Having lerked on the mailing
> list @ linux-ha for a while I can tell you it seems the developers
recommend
> keeping this setting ON. Default is off..
>
> > Please write some of your experiences with heartbeat mon &
> > high availability
> > that could help me!
>
> It's worked good for me for 1.5 years. I use heartbeat + ldirectord + mon
> and have lots of flexibility. Perhaps keepalived (http://keepalived.org)
> might be worth looking into, maybe it fits your needs better. Also since
> you indicated you are using DB's in this setup check out drdb to help keep
> your disks in sync...
I am a little bit confused, becuase you wrote, that you are using heartbeat
+ ldirectord + mon.
I read about ldirectord, that it periodically requests a known URL and
checks the response.
For my business the reply of the watched service would not be a static
output, so I thought, that mon would match better to my requirements.
Why do you use both of this packages? Is this to eliminate the single point
of failure, if you would use only ldirectord?
I think, that keepalived is only for HTTP or other TCP/IP applications, but
I also need to monitor some other things like databases and other thinks,
that I don't know today. Because mon is extremly felxible and easy to use
and configure, I would like to use it ...
DB & Replication:
You wrote something about DB & drdb. I would like to run the database on the
same server as the application does. So I would use the functions
that are build in the databases to replicate them from the primary server to
the backup server over network. So I think, that I don't need to share a
filesystem, isn't it?
One more time thanks to your help
Bye, Gregor
>
> > Thanks, a late merry christmas ;-) and a happy new year (for those who
> > celebrate this ...)
> > Gregor Goestl
>
> Cheers & same,
>
> Peter
>
> _______________________________________________
> LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
> Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
> or go to http://www.in-addr.de/mailman/listinfo/lvs-users
>
>
|
