|
and i (having tried all of these 3 packages discussed), am using heartbeat
for failover of the directors, and keepalived for real server checks. the
curent version of heartbeat is rock solid, though 1 specific problem has
been noted with the latest redhat kernels (which seems to be a kernel
problem, not a problem in heartbeat).
as far as i can see though, keepalived is not a time proven platform for
failover management, i had problems with it, and problems are still reported
from time to time to keepalived list. i prefer using keepalived for real
server checks to using ldirectord, because i am running minimum environment
on the director, so perl is not installed, and keepalived seems to do a fine
job there.
Alex.
----- Original Message -----
From: "Matthew S. Crocker" <matthew@xxxxxxxxxxx>
To: <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Sent: Saturday, November 23, 2002 11:21 PM
Subject: Re: Some thoughts on Ldirectord...and heartbeat
> On Sat, 23 Nov 2002, Malcolm Turnbull wrote:
>
> I know this doesn't really answer your question but you should probably
> switch away from ldirectord and heartbeat. I had the same type of
> problems as you mentioned. I switched to keepalived a couple months ago.
> It *is* a much better software package. keepalived is not bug free but it
> is ready for a production environment.
>
> You can get it at keepalived.sourceforge.net
>
> -Matt
> >
> > I've had a few funnies with this :
> >
> > Sometimes ldirectord will terminate .. Is this due to un-handled signals
?
> >
> > What is the best way to restart on failure ? I'm using an hourly cron
> > job check to restart on failure
> > .. but could sub ld_handler_term be changed to auto restart ldirectord ?
> >
> > Also if it does terminate why would you want to tear down the IPVS table
?
> > I've remarked out the whole of ld_stop to do this, maybee it should be
> > a config option ?
> >
> > Slightly harder to track down but sometimes the ldirectord daemon is
> > still running but seems to stop responding to changes to the
> > ldirectord.cf file ? I'm still working on whether it is checking the
> > servers or not at this stage..
> > I'm restarting it with a daily cron job to get round this .. anyone else
> > had similar problems ?
> >
> > Heartbeat.. probably another mailing list for this but just in case ..
> > sometimes this dies on both the slave and the master server whats the
> > best way of doing an auto restart ? and has any one else had trouble ?
> >
> > IPVS has been solid as a rock however... :-)
> >
> > Regards,
> > Malcolm.
> >
> >
> >
> >
> >
> >
> >
> > _______________________________________________
> > 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
> >
>
> --
> ----------------------------------------------------------------------
> Matthew S. Crocker
> Vice President / Internet Division Email: matthew@xxxxxxxxxxx
> Crocker Communications Phone: (413) 746-2760
> PO BOX 710 Fax: (413) 746-3704
> Greenfield, MA 01302-0710 http://www.crocker.com
> ----------------------------------------------------------------------
>
>
> _______________________________________________
> 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
>
|
