why I get cmcld: failover optimization disabled.

[Ra Jose]

Hello freinds,

I have 3-node and 4-node SG 11.17 cluster,
all nodes on rx8640 running 11.23 with all
latest applied patches.

Whenever I start the cluster or do a cmrunnode,
I keep getting this message in syslog.log file.
This does not repeat though. Only 1 message.

Wondering what this relates to. Tried on ITRC to see any prior posts existed, could not find
any. Also, tried looking at options for cmcld
daemon. My cmcld starts with -j option and not sure what that mean.

ps -ef | grep -i cmcld

shows

root 6527 6495 0 Jul 12 ? 10:28 /usr/lbin/cmcld -j

Is there any connection of cmcld -j with the
"failover optimization disabled". Is there
anyway to enable it.

Thank you for your replies.

Ra Jose.

[Rita C Workman]

Hmm...maybe Stephen or Melvyn can pipe in with the details.

But I have noticed that the primary node tends to show the -j option for the cluster daemon.
While the other nodes that were started after the primary simply show the cluster daemon and the remaining nodes listed.

Just my observation,
Rgrds,
Rita

[Duncan Edmonstone]

Failover optimization is a feature of Serviceguard Extension for Faster Failover:

http://docs.hp.com/en/T2389-90011/ch01s05.html#d0e542

As SGeFF is for 2-node clusters we can safely assume that your system isn't a SGeFF cluster, so just ignore the message.

HTH

Duncan

[melvyn burnard]

Well I do not believe you should be seeing this message.
I would check all nodes to see if they have the SGeFF product installed, and if so remove it as it can only be used for a two node cluster.
Then I woul dverify what patch is installed for SG, and look at installing the latest patch for SG A.11.17 PHSS_37308

[Ra Jose]

Thank you Rita, Duncan and Melvyn. Here is what I have on this 3-node rx8640 IA 11.23
cluster running SG11.17.

root@f1e1pd05:grep -i failover syslog.log
Jul 12 13:12:56 f1e1pd05 cmcld[6694]: Failover Optimization is disabled.
root@f1e1pd05:swlist -l bundle | grep -i "serviceguard"
T1905BA A.11.17.00 Serviceguard
root@f1e1pd05:swlist -l product T1905BA
# Initializing...
# Contacting target "f1e1pd05"...
#
# Target: f1e1pd05:/
#

# T1905BA A.11.17.00 Serviceguard
T1905BA.ServiceGuard A.11.17.00 Serviceguard SD Product
T1905BA.Package-CVM-CFS A.11.17.00 HP SG Cluster CVM CFS SD Product
T1905BA.Package-Manager A.11.17.00 HP Package-Manager SD Product
T1905BA.Cluster-Monitor A.11.17.00 HP Cluster Monitor SD Product
root@f1e1pd05:show_patches|grep -i "serviceguard"
PHKL_35420 Overtemp shutdown / Serviceguard failover
PHSS_35371 Serviceguard A.11.17.00
root@f1e1pd05:what /usr/lbin/cmcld
/usr/lbin/cmcld:
Build platform: hpux
Build date: Tue Sep 12 09:14:20 PDT 2006
Build id: ibld_sg_a1117patch_1123_product
Cluster Monitor Product $Revision: 82.2 $
Cluster Monitor Product Only $Revision: 82.2 $
Daemon
A.11.17.00 Date: 09/12/06 Patch: PHSS_35371

We do not have SGeFF product. What we have is the basic SG 11.17 version. The SG patch we have may not be the latest one, but may be one before the latest.

Does the new patch Melvyn mentioned, has fix for this issue (from patch release notes).

Thank you all,

Ravi

[melvyn burnard]

You need to look for T2389BA as the product and check on ALL nodes.
If it is not there, then you may want to log a call with the HP Response Centre