• iManager Cluster management stopped working with dreaded 'This user does not have the correct credentials to authenticate to the cimom client'
  • Linux workstation objects have blank 'LUM enabled services' window
  • The only tangible issue that I run into from time to time is the CIMOM service is occasionally stopped. Bringing it back up always works. This could be from restarting another service.
  • Assume that cimom issue is from Open WBEM service not being seen as Linux-enabled across all servers.
  • All other services are working wonderfully. Clustered volumes are up and running but cannot be managed except at the command line.


Issue exists on all nodes of a 6-node OES Linux-only cluster. Did the install per the docs:

Installed new OES2Linux nodes into existing NW6.5SP8 eDir tree running latest SP5 of Edir
  • - Did pattern install with necessary services including edir and NCS
  • - Ran updates from NCC at end of install.
  • - created test iSCSI cluster at end of process and everything worked fine.
  • - Left things alone for a while until backup solution (SEP rocks!) was configured.
  • - Had to recreate a number of certs in Netware world using PKIdiag a few weeks later.

At this point it became obvious that I missed a couple of things:
  • non-routable iSCSI interfaces were mapped to eDirectory. Repaired at install time but needed to reinstall NDS on cluster #1 because I didnt RTD. This was a huge learning experience.
  • Didn't create/LUM enable required OES2Linux proxy users at the correct time - pre install - had to create these manually after
  • Put cluster#1's LUM base in one spot and the rest in another. This was rectified at the reinstall of cluster#1 into the tree with the deletion and recreation of the servers LUM objects in edirectory.

Steps taken during this process:
  • Fixed the iSCSI interface issue by modifying the NDS configuration file and hosts file per TID's. Repaired server addresses and everything worked perfectly.
  • Fixed the proxy users issue during the edir repair process.
  • Removed edir from cluster#1 and reinstalled it after removing all objects from edir.
  • Recreated and reconfigured LUM per numerous TID's so that all tests such as ID user work fine.
  • Changed LDAP mappings for uniqueID per the OES2 LUM troubleshooting docs
  • Did numerous LDAP traces to make sure that there were no errors
  • Checked openwbem conf file and NAM.conf files for problems. Rebuilt all of them to make sure.

Dumb things I did that may have created this issue:
  • Used ConsoleOne to manually edit fields in LUM objects using 'other' field. This is supposed to be OK ....if you know what you are doing
  • Split my LUM bases up during install by forgetting to add objects to same place.
  • Attempted to reconfigure eDir using YAST after eDir already installed on one server. Needed to do manual removal and reinstall per TIDS.
  • Got admin rights warning during LUM install on some servers (object not Linux enabled) and didn't follow up on it .

Tests performed during head scratching process:
  • DS repairs are all completely fine - better than ever in fact from all servers
  • Did LDAP traces to check for auth issues with LUM. No errors evident.
  • Looked at private test cluster created before the install. Test cluster was flawless.
  • Compared this production install with test install config files and objects. All the same.
  • ID command and NAM listings all come up correctly
  • Startup logs don't show any NAM errors
  • Re-configured/verified LUM groups, Linux config object and Linux workstation objects
  • Set login for OPENWBEM/CIMOM to allow anonymous login

Potential options on table now that I've had it - PLEASE ADVISE!!!
  • Remove all LUM objects from tree and reinstall/reconfig LUM via YAST
  • Remove all servers in cluster and start over. A shame since all else is working but the cluster isn't in production yet so it's an option.
  • Advice from Novell Forums
  • Burn an Incident

Many thanks!