MTA Shows Other MTAs as Closed, Link or Transport Down

  • 7003952
  • 22-Jul-2009
  • 27-Apr-2012

Environment

Novell GroupWise 8 HP2
Novell Open Enterprise Server 2 (OES 2) Linux SP1
NSS File System

Situation

GroupWise System is installed on multiple clusters on machines running OES2 SP/NSS.

The primary domain and several secondary domains are running in one of the clusters. The primary domain MTA shows all of the secondary domain MTAs (in the same cluster as the primary) as closed; Last Closure Reason is "Link or Transport Down".

However, all the secondary MTAs show the primary MTA as open. Also, primary MTA shows all secondaries in other clusters as open.

Resolution

Restarting primary MTA did not resolve problem.
Restarting primary and secondary MTAs did not resolve the problem.
Unloading and reloading primary MTA did not resolve the problem.
Rebuilding primary domain database did not resolve the problem.
Validating primary domain database did not resolve the problem.
Verified that customer is using Bind Exclusive instead of local IP address on cluster.
Renaming the /wpcsin, /wpcsout, and /mslocal directories did not resolve the problem.
However, we saw some strange behavior after renaming the directories. After renaming all of the queues and restarting the MTA the /mslocal directory always had the same files in it.
We tried deleting all of the message queues (including the ones we had renamed) and restarting the MTA. This did not resolve the problem.
Finally, we renamed the domain directory, created a new domain directory by the same name, copied the wpdomain.db and all of the .dc files and restarted the MTA. This resolved the problem.
We suspect that there is filesystem corruption.