[Pacemaker] Pacemaker not starting in 2nd node of a 2 node cluster

[Enric Muñoz]

I only see the multicast packets in the node that is online but not in the other one. How can I configure corosync to use unicast??

After enabling IGMP in the switch it appears a new message in the LCMC:

Could not initialize corosync configuration API error 6
Crm verify: Could not establish cib_ro connection: connection refused (111)
Live CIB query failed: Transport endpoint is not Connected.
Connection to cluster failed: transport endpoint is not Connected.

Thanks

De: Michael Schwartzkopff [mailto:misch at clusterbau.com]
Enviado el: lunes, 29 de julio de 2013 12:31
Para: The Pacemaker cluster resource manager
Asunto: Re: [Pacemaker] Pacemaker not starting in 2nd node of a 2 node cluster


Am Montag, 29. Juli 2013, 10:06:29 schrieb Enric Muñoz:

> Iptables is disabled and selinux set to permissive in both nodes.



Do a tcpdump on both nodes. You should see the multicast packets as well as some unicast packets to the udp port you defined. As fasr as I remember there are three (or four?) unicast packets between any multicast packet.

If you do not see these packets, you have several options to search:



- do you use virtual machines? I have seen kvm problems with multicast packets.

- separate subnets? Talk to your network guys to use IGMP.

- Problems with your switches? Talk to the network guys.



On the other hand, you also could use udpu (unicast) for the transport.



--

Dr. Michael Schwartzkopff

Guardinistr. 63

81375 München



Tel: (0163) 172 50 98
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.clusterlabs.org/pipermail/pacemaker/attachments/20130729/e52dbb2d/attachment.htm>