Second node can't see one shared disk

[mikerault]

I have a 2-node setup using 32 bit servers running Redhat Linux 2.6.9-67.0.15 connected through a brocade switch to twho disk arrays, a NStor 18f and an NStor 8f. Node 1 sees all 26 drives, node 2 sees all but the last disk on the 18f.

During the POST the fibre card shows that it sees all 26 drives but dmesg output shows that the last drive is never mapped to /dev/sdz as it should be.

Any ideas out there?


Mike

 

[LiamC]

Can you clarify exactly what type of nodes you are talking about.

You haven't set the the upper bound and forgotten to leave space for one extra (DLM_LOCK_RES_OWNER_UNKNOWN and such) have you?

 

[mikerault]

I am using the Oracle CRS to cluster the two nodes for use with an Oracle11g database. Oracle handles the DLM settings internally.

Mike

 

[mikerault]

he problem was that the FC909 HBA wouldn't allow reset of the host address, the interface would allow you to change it to 27 from 15 on the screen but then on reboot would reset it to 15 which then caused issues. I replaced the FC909 based HBA with a qla2200 and the problem was resolved.

Mike