Having trouble with getting VCS 'had' to start.
Created: 13 Nov 2012 | Updated: 14 Nov 2012 | 7 comments
This issue has been solved. See solution.
I installed VCS 6.0.2 on RHEL 6.1, chose not to start LLT and GAB during configuration so that it would be configured a single node cluster in a standalone cluster.
However 'had' failed to startup, and the error message at the end of installation was as follows :
"It is likely that the startup failure issues will be resolved after rebooting the system. If issues persist after reboot, contact Symantec technical support or refer to installation guide for further troubleshooting."
I restarted the system, but unfortunately 'had' still would not start after running 'hastart'.
The file /var/VRTSvcs/log/engine_A.log does indicate that 'had' started :
2012/11/13 21:24:27 VCS NOTICE V-16-1-11022 VCS engine (had) started
2012/11/13 21:24:27 VCS NOTICE V-16-1-11050 VCS engine version=6.0
2012/11/13 21:24:27 VCS NOTICE V-16-1-11051 VCS engine join version=6.0.10.0
2012/11/13 21:24:27 VCS NOTICE V-16-1-11052 VCS engine pstamp=6.0.100.000-GA-2012-07-20-16.30.01
2012/11/13 21:24:27 VCS NOTICE V-16-1-10114 Opening GAB library
2012/11/13 21:24:27 VCS INFO V-16-1-10196 Cluster logger started
but, a 'ps -ef | grep had' returns no entry.
The file /var/VRTSvcs/log/hastart.log has this entry for the startup attempt :
File was modified after system boot,
boottime = <1352828351>, modifytime = <1352828394>
[root@localhost log]# view engine_A.log
Discussion Filed Under:
Comments 7 Comments • Jump to latest comment
Can you try "hastart -onenode" ?
If you have not configured LLT and GAB, then to start VCS you must use:
and "hastart" on its own without "-onenode" flag will not work with GAB.
This should work on bootup as the RC scripts have something like:
if [ "$gabconfigured" = "Configured" ]; then $HASTART else $HASTART -onenode fiOne additional thing to check is that you have "ONENODE=yes" inside the file /etc/sysconfig/vcs.
Mike
UK Symantec Consultant in VCS, GCO, SF, VVR, VxAT on Solaris, AIX, HP-ux, Linux & Windows
If this post has helped you, please vote or mark as solution
Thanks very much Tushar and mikebounds for your suggestions.
I tried 'hastart -onenode', but unfortunately 'had' is stiill not in the list of processes :
I wonder if the problem is license-related. I am using 60-day trialware which I originally downloaded and installed on the same machine back in Jan/Feb 2012. Shortly after that, I was tasked with something else, and I've only recently been reassigned to invesiigate VCS use in a one-node cluster. I did a fresh install of RHEL6 and of VCS 6.0.2, but perhaps the original installation is flagged as license expired based on the MAC address on my NIC?
Then again maybe I wouldn't have gotten that far with the install if there was a licensing issue?..
I don't think issue is licencing as this should show in engine log - something "VCS is not licensed" and with 6.0, it comes with its own 60 day eval, but even after 60 days, VCS will still work, you will just get warnings in the log. You can verify if you have the built-in keyless 60 day eval by using:
Can you provide the following:
Output from command "gabconfig -a" (expecting it to say nothing)
Contents of main.cf (in /etc/VRTSvcs/conf/config)
Usually if "had" can't start it says in the engine log why it can't so it looks like "had" is just dying rather than exiting.
Mike
UK Symantec Consultant in VCS, GCO, SF, VVR, VxAT on Solaris, AIX, HP-ux, Linux & Windows
If this post has helped you, please vote or mark as solution
Have a look at http://www.symantec.com/business/support/index?page=content&id=TECH77211
This basically says "The cause of the GAB driver configuration issue is inconsistency in system name in VCS configuration", so you need a proper name for the host and do not use "system localhost" in your main.cf
Mike
UK Symantec Consultant in VCS, GCO, SF, VVR, VxAT on Solaris, AIX, HP-ux, Linux & Windows
If this post has helped you, please vote or mark as solution
Hi Mike,
Thanks again for your prompt response, and insightful troubleshooting.
I changed hostname from localhost to vcs-node1 in files 'main.cf', '/etc/sysconfig/network' and '/etc/hosts', and rebooted the server.
It appears to have resolved the issue :
Would you like to reply?
Login or Register to post your comment.