Unable to stat /etc/cmcluster/cmclconf

[Ludovic Derlyn]

Hi,

I have a cluster Sap with two nodes and two AS on two others nodes
On this As, i have this message "Unable to stat /etc/cmcluster/cmclconf
ig, No such file or directory" every two minuts

Service guard is installed, but not started.

swlist -l product |grep -i guard
PHSS_33840 1.0 Serviceguard A.11.17.00
SG-Apache-Tool B.03.00 Serviceguard Apache Script Templates
SG-NFS-Tool A.11.23.04 Serviceguard NFS Script Templates
SG-Oracle-Tool B.03.00 Serviceguard Oracle Script Templates
SG-Samba-Tool B.03.00 Serviceguard Samba Script Templates
SG-Tomcat-Tool B.03.00 Serviceguard Tomcat Script Templates
ServiceGuard A.11.17.00 Serviceguard SD Product
in /etc/cmcluster, I have cfs
ll /etc/cmcluster/cfs/
total 192
-r-x------ 1 bin bin 3183 Nov 9 2005 SG-CFS-pkg-preshutdowh
-r-x------ 1 bin bin 2218 Nov 9 2005 SG-CFS-pkg.conf
-r-x------ 1 bin bin 5930 Nov 9 2005 SG-CFS-pkg.sh
-r-x------ 1 bin bin 4288 Nov 9 2005 SG-CFS-sgcvmd.sh
-r-x------ 1 bin bin 22182 Nov 9 2005 SG-CFS-util.sh
-r-x------ 1 bin bin 3583 Nov 9 2005 SG-CFS-vxconfigd.sh
-r-x------ 1 bin bin 2362 Nov 9 2005 SG-CFS-vxfsckd.sh
-r-x------ 1 bin bin 4133 Nov 9 2005 cmvxsw
-r-x------ 1 bin bin 11107 Nov 9 2005 vx-modules

what is this message, and who is the source ?

regards

Ludovic

[Luk Vandenbussche]

Ludovic,

Does the file /etc/cmcluster/cmclconf exist on your server?

Try

cmclgetconf -v -c clustername /etc/cmcluster/cmclconf

[Ludovic Derlyn]

hi,

I didn't exist :

ll /etc/cmcluster
total 32
drwxr-xr-x 2 bin bin 8192 Jan 18 07:06 cfs
-r-------- 1 bin bin 70 Sep 23 2005 cmknowncmds

regards

[Enrico P.]

Hi Ludovic,
do you have any software like openview that is monitoring the system?
Seems it search for the standard configuration cluster file and not find it.

Enrico

[Ludovic Derlyn]

hi,

Yes i use openview for monitoring the system
I will check this.

regards

[melvyn burnard]

a server somewhere has a process that is trying to interrogate Serviceguard nodes.
This is nornmally some type of monitoring software, and the message is because the server being probed has SG installed, but no binary exists.

Look for processes on this server that may be causing thios.

[Stephen Doud]

To extend what Melvyn wrote, the 'cmgetconf' command is the most likely suspect. The following have been observed to perform cmgetconf on servers installed with Serviceguard:
1) Ignite/UX - make_recovery gathers cluster information
2) Legato (Networker) backup software running on the node.
3) cron
4) Serviceguard Manager
5) Predictive HA configuration (if old O/S)
6) HA Tracker tool (used with CSS support and an HAO account)
7) ISEE tool. ISEE replaces the HAO utility.
8) In one case, ovconfd has been found to run cmgetconf every 10 seconds.

You may benefit from this document:
UMCSGKBRC00014697
TITLE: Tracing the source of cmgetconf commands

[Ludovic Derlyn]

Hi,

I use Serviceguard manager , and i thinks that problems occurs by this software.

When I launch Serviceguard manager , Two nodes with this messages appears in state unused

for stephane, I didn't have access to document:UMCSGKBRC00014697
TITLE: Tracing the source of cmgetconf commands, could you send me

thanks

[Chauhan Amit · Accepted solution]

Check out this thread as well

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=994303&admit=-682735245+1150193719526+28353475

Hope this helps.

-Amit

[Ludovic Derlyn]

modifications in inetd.conf resolve this problem

thanks