Suncluster 3.2 on Vmware panic problem

Question

Hello All,
This issue might have been discussed already, but looking for any known solution from anyone.
I have installed Sun Cluster 3.2 on VMware. (Hardware AMD 64 X2, 4G RAM). Everything went on smooth. The cluster is installed configured and quorum configured (shared disk by a vmdk file). Suddenly, it began to get errors as below and nodes started panicing.
Dec 13 05:33:53 vmnode5 genunix: [ID 313806 kern.notice] NOTICE: pm_tick delay of 15281 ms exceeds 2147 ms
Dec 13 05:33:53 vmnode5 genunix: [ID 313806 kern.notice] NOTICE: pm_tick delay of 15267 ms exceeds 2147 ms
Dec 13 05:33:53 vmnode5 genunix: [ID 313806 kern.notice] NOTICE: pm_tick delay of 15280 ms exceeds 2147 ms
Dec 13 05:33:53 vmnode5 genunix: [ID 313806 kern.notice] NOTICE: pm_tick delay of 15266 ms exceeds 2147 ms
Dec 13 05:33:53 vmnode5 genunix: [ID 313806 kern.notice] NOTICE: pm_tick delay of 15267 ms exceeds 2147 ms
I have searched and even updated the patches to 120012-14. Still no luck. Please advice if anyone knows solution for this.
Regards
LP

Answer 1

Thanks for posting this. It's interesting that you got this to work.
The rules for what is used for quorum devices (QDs) and other shared devices are as follows:
_2 paths to the device_
QD: SCSI-2, with PGRe.
non-QDs: SCSI-2 (*no* PGRe is needed for shared devices that are not QDs).
[Ie, PGRe is only needed for QDs.]
_2 paths to the device_
QDs and non-QDs: SCSI-3 PGR
PGRE = persistent group reservation emulation.
So SCSI-3 should only be needed if you have a >2 node cluster or more than 2 paths to your storage. If you use MPxIO then the dual paths would be hidden.
So, in theory, you should be able to try out RAC.
Let us know how you get on!
Tim
---