Cluster not operational when gab is started there is no LLT communication.

Article:TECH201928  |  Created: 2013-01-22  |  Updated: 2013-01-22  |  Article URL http://www.symantec.com/docs/TECH201928
Article Type
Technical Solution


Environment

Issue



After a SAN(Storage Area Network) outage and reboot of cluster nodes. GAB (Global Atomic Broadcast) is unable to get cluster membership. The LLT links which are connected via crossover cable are no longer able to successfully transmit packets to each other, which means no GAB membership. Customer is able to ping the interfaces and verify success of LLT links, but enabling GAB results in numerous LLT link errors and no node membership.


Error



gabconfig -c

Results in a flood of LLT messages

LLT INFO V-14-1-10205 link 0 (eth4) node 1 in trouble
LLT INFO V-14-1-10024 link 0 (eth4) node 1 active
LLT INFO V-14-1-10205 link 1 (eth5) node 1 in trouble
LLT INFO V-14-1-10024 link 1 (eth5) node 1 active
LLT INFO V-14-1-10205 link 0 (eth4) node 1 in trouble
LLT INFO V-14-1-10024 link 0 (eth4) node 1 active
LLT INFO V-14-1-10205 link 1 (eth5) node 1 in trouble
LLT INFO V-14-1-10024 link 1 (eth5) node 1 active
 


Environment



RHEL5U8
Storage Foundation HA 6.x
HP Blade
Hewlett-Packard Company NC325m PCIe Quad Port Adapter
 


Cause



The issue is with the frame the blade was located in. There is a bug that causes issues with packets 1500 bytes or larger in the server enclosures.

With the help of there hardware team they can test for the bug by doing a

# ping -L 1500

If it works without the "-L1500 flag", then you have not hit the bug
but if does not with it, it's more than likely the bug.

 


Solution



Run the "ping -L 1500" test to verify if the frame can handle packets greater than 1500bytes. If it cannot contact the hardware vendor.




Article URL http://www.symantec.com/docs/TECH201928


Terms of use for this information are found in Legal Notices