VCS Global Clustering - WAC Error V-16-1-10543 IpmServer::open Cannot create socket errno = 97

[Rich_P]

Hi,

I have a customer who has two VCS clusters running on RHEL 5.6 servers. These clusters are further protected by site failover using GCO (global cluster option). All was working fine since installation with remote cluster operations showing up on the local cluster etc. But then this error started to appear in the wac_A.log file ....

VCS WARNING V-16-1-10543 IpmServer::open Cannot create socket errno = 97

Since this the cluster will not see of the remote clusters state, but can ping it as seen from the hastatus command below:

 

site-ab04# hastatus -sum

-- SYSTEM STATE
-- System               State                Frozen

A  site-ab04       RUNNING              0

-- GROUP STATE
-- Group           System               Probed     AutoDisabled    State

B  ClusterService  site-ab04       Y          N               ONLINE
B  SG_commonsg     site-ab04       Y          N               ONLINE
B  SG_site-b04g3 site-ab04       Y          N               OFFLINE
B  SG_site-b04g4 site-ab04       Y          N               OFFLINE
B  SG_site-a04g0 site-ab04       Y          N               OFFLINE
B  SG_site-a04g1 site-ab04       Y          N               OFFLINE
B  SG_site-a04g2 site-ab04       Y          N               OFFLINE
B  vxfen           site-ab04       Y          N               ONLINE

-- WAN HEARTBEAT STATE
-- Heartbeat       To                   State

M  Icmp            site-b04c       ALIVE

-- REMOTE CLUSTER STATE
-- Cluster         State

N  site-b04c  INIT

Does any one have any ideas? Networking all seems to be in order.

Thanks,

Rich

[mikebounds]

Are you using secure clusters (i.e is the cluster attribute which you can view in main.cf or from "haclus -display" set to 1) - if so then the issue could be as in http://www.symantec.com/docs/TECH152575

Mike

[Rich_P]

Hi Mike,

No its isn't secure clusters. IPv6 is disabled but I am running the commands as root. Also we have another identical cluster that is running ok.

Rich

[Pranav_Sarwate]

Hi Rich,

Can you provide the ping output between the machines to see what is the approx delay in network? Also, can you provide the ConnectTimeout value for the cluster? It's possible that the network delay is more than the connection timeout in VCS, due to which the remote cluster connection remains in the INIT state.

- Pranav