LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: ipvs failback patch

To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Subject: Re: ipvs failback patch
From: Horms <horms@xxxxxxxxxxxx>
Date: Tue, 22 Nov 2005 06:58:43 +0000 (UTC)
Ranga Nathan <kairanga@xxxxxxx> wrote:
>>ipvsadm --start-daemon master
>>ipvsadm --start-daemon backup
>>  
>>
> I am sorry, I did not explain very well. My language is not very 
> technical :-) I did  what you suggested before. From master to backup 
> the failover worked fine. I did not lose any connections. From backup, 
> when the master (after a reboot) snatched back the nodes, the 
> connections dropped.  I am sure that when the master came backup, it 
> started the daemon, as I had "--start-daemon master" in the 
> /etc/ipvsadm.rules. I confirmed this by doing "ipvsadm --start-daemon 
> master" on the master and it said "Daemon has already run".  I could not 
> query if it is in "master" status. I presumed so.

What is the output of ps ax | grep syn. If you don't seem two entries,
one for syncmaster and one for syncslave, then connections will
only ever be synchronised one way. If you see two (on both nodes)
then connections will be synchronised both ways.

> I there a way for the master and backup swap dynamically when snatching 
> the nodes?

The basic problem is that when the old-master reboots,
if it snatches back master status straight away, then
there won't have been time for the other node to syncrhonise
connections as packets for them are recieved. What you 
really need to do is to be able to have a node ask the
other node to send over all the information it has,
and to know when that is complete.

I'm not aware of any patches that do that, though that
doesn't mean they aren't out there. However, an alternate
is to make the rebooted node be a stand-by until such
time as the other node fails.

-- 
Horms


<Prev in Thread] Current Thread [Next in Thread>