heatbeat ERROR: NV failure (msgfromsteam): [>>> ] ha_msg_add_nv_depth: line doesn't contain '='
Apr 17 11:55:47 Cluster01 heartbeat: [1426]: WARN: ha_msg_add_nv_depth: line doesn't contain '='
Apr 17 11:55:47 Cluster01 heartbeat: [1426]: info: >>>
Apr 17 11:55:47 Cluster01 heartbeat: [1426]: ERROR: NV failure (msgfromsteam): [>>>
]
I have never been able to reproduce the above but it happened on heartbeat-3.04-2 when I took a node down for testing.
One thing I believe caused it was by configuring ha.cf to do bcast br0 (which means all other heartbeat nodes on the network will communicate when they shouldn't). I verified this by fluke when I was shocked to see communication from the bad heartbeat node with other unrelated nodes.
I believe that is what introduced the above errors so I switched to ucast and specific IPs only rather than bcast.
Tags:
heatbeat, nv, msgfromsteam, ha_msg_add_nv_depth, doesn, contain, apr, cluster, heartbeat, info, reproduce, node, configuring, cf, bcast, br, nodes, communicate, shouldn, verified, fluke, communication, unrelated, errors, ucast, ips,