cmcheckconf hangs

[Dan Copeland]

Admins,

I tried running a cmcheckconf -C file -k -v and it has been running for 6 hours! Some times they take a while (six node cluster) so I let it run and went home. I checked it now and it's still running. I'm creating the cluster for the first time but the cmquerycl was hanging too, so I built the ascii file myself and tried the cmcheckconf.

The only thing in syslog is a warning about not being able to stat /etc/cmcluster/cmclconfig but it shouldn't because there is no cluster yet, correct?

Any thoughts on why it would hang?

tia,
Frank

[Helen French]

I would first think why the cmquerycl was hanging? Did you check the network and it's configurations? Just go through this document and make sure that you didn't miss anything while configuring the cluster (TKB #UMCSGKBRC00010342):
http://www1.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000063198778

Also, which version of SG are you using? which OS? Did you install all patches before the configuration?

[Dan Copeland]

I apologize for not searching the archives, but I'm connected at 24 kbs and it's 2 am. I saw the post a few below this one will check for potential disk errors and proper resolution. Any other thoughs?

thanks,
Frank

[Dan Copeland]

MC/SG v. 11.13

Did linkloop tests for all scenarios--worked fine.

[Ashwani Kashyap]

Check for disk errors or network errors or configuration .

I have seen it usually hanging on either of the two .

[Helen French]

Make sure that you have this patch installed if you are running 11.X - PHSS_27087

[linuxfan]

Hi Frank,

I have seen cmcheckconf hang if you don't have /etc/cmcluster/cmclnodelist or /.rhosts setup.

Check
http://docs.hp.com/cgi-bin/fsearch/framedisplay?top=/hpux/onlinedocs/B3936-90053/B3936-90053_top.html&con=/hpux/onlinedocs/B3936-90053/00/00/32-con.html&toc=/hpux/onlinedocs/B3936-90053/00/00/32-toc.html&searchterms=cmclnodelist&queryid=20021003-212130

The setup of cmclnodelist is explained in the section Preparing your System (5. Building an HA Cluster Configuration)

You have to setup the cmclnodelist on all the nodes that you are adding to the cluster.

-HTH
Ramesh

[Dan Copeland]

The .rhosts were setup, but grasping for straws, I setup the cmclnodelist as well. I don't think this is the issue. What gives the most info is a warning I get when I run: cmquerycl -l lvm -v -k -n node1

Warning: No possible heartbeat networks found.
All nodes must be connected to at least one common network. This may be due to DLPI not being installed.

This is also strange since I'm getting this when trying w/ only one node...What's the DLPI all about?

tia,
Frank

[Frederic Sevestre]

Hi Franck,

I sounds like the problem I just have. Check this link :

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x6c83402f24d5d61190050090279cd0f9,00.html

Regards,
Fr??d??ric

[Dietmar Konermann]

Hi, Frank!

Don't worry about the DLPI warning... it's quite normal if you specify -l lvm (looks like a bug for me).

But indeed the thread Frederic pointed to should give you hints how to troubleshoot this...

Use the -T 5 opion to get cmclconfd's debug logging enabled. If this does not help ese tusc and q4 to have a closer look at it.

Regards...
Dietmar.