LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Re: ipvs failback patch

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: ipvs failback patch
From: Ranga Nathan <kairanga@xxxxxxx>
Date: Mon, 21 Nov 2005 20:54:01 -0800


Horms wrote:

Ranga Nathan <kairanga@xxxxxxx> wrote:
Horms wrote:

Ranga Nathan <kairanga@xxxxxxx> wrote:


Karl Kopper's Linux Enterprise Cluster talks about the availability of a kernel patch to failback from the backup to the master that keeps the connections alive. I could not find it on the linuxvirtualserver.org site. I see the latest versions but they are in source format. Somehow I am unable to do rpm --rebuild <sourcerpm>

The current version I have is: ipvsadm v1.24 2003/06/07 (compiled with getopt_long and IPVS v1.2.0)
Could you be a little more specific about the patch
(or at least give a page refernce in the book).


It is the NOTE at the bottom of p276 (Stateful Failover of the IPVS Table)

I am suspecting that this patch is no longer needed.


When it fails over, the connections are retained. When it fails back, the connections are broken.
I simulated this by rebooting (this is my lab machine!)  the directors.

Ok, I'm still a little confused, but I think what he is talking
about is being able to run the master and backup daemons on
both linux-directors at the same time. That patch was merged
into the kernel a while ago now, and the chances are that
you can do this without patching. To test, try running:

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.

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

Which if it works, will improve things somewhat.


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