BU050412_EW01

Tru64 UNIX

Evolving business value

Related links

Loss of access to EVA storage units can occur in a Tru64 UNIX cluster after a system update to V5.1B-2/PK4. This problem occurs only on some nodes of the cluster, while other nodes continue to have access to the units.

This problem does not cause data loss or corruption. It is limited only to direct data access from some nodes of a TruCluster system.

Cause

This problem is due to incorrect application of the 7-November 2003 -- Engineering Advisory (EA) number BU030929_EW01.

If that EA is applied incorrectly (or to inappropriate versions of the Tru64 operating system) and the system is later upgraded to V5.1B-2/PK4 or higher, it might exhibit the loss of access problem.

Note: The symptom is related to EVA storage, but is caused by an incorrect MSA entry in the /etc/ddr.dbase. You must correct this entry even if MSA storage is not connected to the system.

The problem has only been reported after clustered systems are upgraded to V5.1B-2/PK4 using the rolling upgrade process.

Please check all nodes of the cluster to determine whether or not they have this problem.

Run the following check procedure independently on each node of the cluster. Only if you detect the problem, please elevate an L3 case for assistance.

Login as root user and run the following command:

# ddr_config -c

The command is successful if the ddr.dbase file compiles cleanly, and the prompt is returned with no error.

If this command succeeds, the ddr.dbase file is not corrupted and this EA does not apply.

If the command produces the following error (the line number may be different), the system has the problem:

# Error: Line 352: parameter already specified.

In this situation, please elevate an L3 case for assistance in correcting the database.