Unable to start cluster and receive GAB error : Module 7, Line 389, Error c000000d
|Article:TECH171221|||||Created: 2011-10-06|||||Updated: 2012-04-06|||||Article URL http://www.symantec.com/docs/TECH171221|
All attempts to start the cluster ends in error.
System Event Error:
Event Type: Error
Event Source: GAB
Event ID: 16
GAB: Module 7, Line 389, Error c000000d
# for hex 0xc000000d / decimal -1073741811 :
SFW-HA 5.0 RP2, Windows 2003 enterprise R2, 32 bit
Systems in a cluster are connected by redundant cluster communication links. Each cluster has a unique cluster ID and can be specified between 0 and 255. The issue was identified where the cluster ID for the local cluster was not unique. This will cause a number of issues including duplicate configuration on both clusters in the environment and may result in a number of MAC addresses being re-used. This will make the cluster unstable.
This issue can be resolved by following the steps below:
1. Identifiy both clusters that has identical cluster IDs
2. Stop LLT, GAB and VCSCOMM services on all nodes on both clusters 1 and 2
- From a command prompt run the following command: Net stop llt /y
3. Decide on unique numbers for the cluster ID for cluster 1 (eg 100) and a cluster ID for cluster 2 eg (101)
4. Update/Change the clusterID in the llttab.txt file for all cluster nodes on both clusters 1 and cluster 2
5. Make sure llt is stopped on all nodes in both clusters 1 and 2
6. When nodes restart, Start llt, gab, and vcscomm on one of the cluster nodes on cluster 1
7. Once all 3 components are restarted (with the new clusterID in the llttab.txt), run gabconfig -a on at least one node in cluster 1 to ensure you see port membership for port a. Start the cluster on that same local node by running hastart
8. Start cluster on cluster 1 by running hastart - all and make sure all start up successfully.
9. Start llt, gab, and vcscomm on one of the cluster nodes on cluster 2
10. Once all 3 components are restarted (with the new clusterid in the llttab.txt), run gabconfig -a on at least one node in cluster 1 to ensure you see port membership for port a. Start the cluster on that same local node by running hastart
11. Start nodes on cluster 2by running hastart - all and make sure all start up successfully.
Article URL http://www.symantec.com/docs/TECH171221