On 12/14/2012 07:22 AM, Muhammad Sharfuddin wrote: > On Fri, 2012-12-14 at 14:09 +0100, Fabian Herschel wrote: >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA1 >> >> Hi Muhammad, >> >> also find my ansawer inline... >> >> On 12/14/2012 12:55 PM, Muhammad Sharfuddin wrote: >>> On Fri, 2012-12-14 at 16:47 +0500, Muhammad Sharfuddin wrote: >>>> please find me replies in-line >>>> >>>> On Fri, 2012-12-14 at 12:13 +0100, Fabian Herschel wrote: >>>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 >>>>> >>>>> As you are using Multcast (MCAST) >>>>> >>>> yes >>>> >>> so using Unicast instead of MCAST, would be a solution ? >>> >> >> It COULD be a solution, if the network was the problem. >> > Ok, will try it. > But as this problem is not **frequent**, I mean this cluster was running > almost since 2 months without any error, therefore after switching to > Unicast I have to wait for at least next 2 months to be sure that using > Unicast is a fix. > >> Some years ago I wrote a tiny programm to send MCAST with >> high load and to count drops - maybe I can reload the code >> and sent it to you, if you are interested. Its GPL so free >> to use under the terms of GPL and no warranty :) >> >> > Waiting.. thanks, appreciated a lot.
Cisco switch? I've heard of them (possibly others) periodically drop multicast groups as a way to clean up stale groups, figuring the client will just re-request membership. Trouble is, it takes too long so the cluster fails. The answer is to set a static multicast group in the switch. I believe you can use omping to test multicast. -- Digimer Papers and Projects: https://alteeve.ca/w/ What if the cure for cancer is trapped in the mind of a person without access to education? _______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
