Packages configuration disappeared on MC/SG 11.16 !!!

[Romaric Guilloud]

Hello there,
I have a quite critical issue with a MC/SG 11.16 cluster on which nodes I installed PHSS_32731 (along with its dependencies) following a call open at HP support:
Now, the 3 nodes are patched and when reforming the cluster, the packages configuration HAS DISAPPEARED!
cmviewcl only shows nodes as up and running and there is now mention AT ALL of my packages.
Where is the package config. gone?
cmapplyconf doesn't work and claims "Failed to lookup /cluster in configuration database".
Your help is greatly welcome out there.
Thanks in advance.

Romaric.

[Kent Ostby]

Romaric --

Did you rebuild the config file from scratch during the process at all ? If you did (i.e. a cmdeleteconf and then cmcheckconf/cmapplyconf), than it may be that you didn't include the packages in the cmcheckconf / cmapplyconf.

Please post cmviewcl output.

Please post cmcheckconf output as well.

[Romaric Guilloud]

Hello Kent and thanks for your quick feedback.
I didn't do cmdeleteconf before patching my nodes, just stopped the nodes, patchedm the reformed the cluster.
Here is the output of cmviewcl:

glen2 # cmviewcl

CLUSTER STATUS
glen up

NODE STATUS STATE
glen1 up running
glen2 up running
glen3 up running
glen2 #


All packages ARE GONE!

Then cmcheckconf of a given package returns:
glen2 # cmcheckconf -v -P ./efr1/dbefr1.ascii

Checking existing configuration ... Done
Error: Failed to lookup /cluster in configuration database.
Error: Fail to load data from configuration database.
Internal error: Unable to open communications to configuration daemon: No such file or directory
cmcheckconf : Cluster is not currently configured. -C option is required.
glen2 #


Any idea of what is going on?
Rgds,

Romaric.

[Romaric Guilloud]

Hello Kent I finally had the answer from HP support, it was about adding -i option to cmclconfd in inetd.conf (new thing to do with MC/SG 11.16).
Thanks anyhow for your time.

[Stephen Doud]

I don't recommend that you continue to use the '-i' workaround to keep your cluster running.
PLEASE review the Special Installation Instructions in the patch to learn what system configuration changes you need to make to permit the use of 'identd' daemon to secure Serviceguard further.

[Romaric Guilloud]

Right Stephen now that all nodes are at the same patch level, I did re-enable identd.
Rgds,