LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

ultramonkey heartbeat and ldirectord problem

To: "LinuxVirtualServer.org users mailing list." <lvs-users@xxxxxxxxxxxxxxxxxxxxxx>
Subject: ultramonkey heartbeat and ldirectord problem
From: Charles Holbrook <cholbrook@xxxxxxxxxxxxxx>
Date: 24 Jun 2003 15:53:22 -0500
SHORT HISTORY:  atm1 server tanked last week when processor fan melted
into heatsink.

Pulled system from production pulled harddrive and put it into new
system.  kudzu found all hardware and configured it correctly.  an fsck
found no errors on the drive whatsoever, data and conf files had same
sums as before meltdown.

Put atm1 back in to production,  atm2 relinquished control and atm1 took
over again.  HOWEVER...

when atm1 takes over it will create the virtual IP addresses bound to
the correct interfaces but it will not actually pass any traffic.  An
ipvsadm shows the number of inactive connections to be incrementing the
exact value of the number of attempts I am making through it but no
connection ever returns any information.  Logs show that about 10
seconds after binding the virtual IP addresses they are unbound but
still show up in an "ip addr list."

Current config files have the same sum value for ldirectord.cf and
ha.d.  Does anyone have any ideas what could of caused this to happen. 
I know most of you will say why not just replace the drive and see if
that fixes the problem.  Well no cash and drive looks good under
PCCertify's drive test as well.

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