MCSG 11.09: EMC 8730 Array & Two Node Cluster

[LG Porter]

I have a two cluster composed of two K580 class servers in a MCSG two node cluster configuration with a EMC storage array. New EMC disks have been added to the storage pool, what is the best method of identifying EMC "DEV" number and "LUN's" with the primary and the failover server? EMC has a utility called "inq" that can be used to identify EMC disks and disk serial numbers, can this also be used?

[hpuxrox · Accepted solution]

Steps to make drives available from an EMC frame,

1. Make available to interfaces with SDR and Volume Logix

2. run ioscan

3. run insf -e

4. Confirm drives with ioscan and inq

[Animesh Chakraborty]

Hi,
Yes.inq can be used to identify the disk for both the nodes.

[Roger Baptiste]

hi,

Since this is a cluster, make sure both the nodes are exactly identical i.e they have the fibre,lan and other I/O cares on similarly numbered slots. Do an ioscan -nf to make sure about it. If it is not identical, this is the best opportunity to correct it. Later, it will be a mess.

Once you do this, get on to the EMC part. If EMC is already connected to the systems (through switch or directly ) and configured to be made visible (through SDR or volumelogix), move to the next step.

On both systems run:
insf -e (to add disk device files)
ioscan -nf >ioscan.out
ioscan -nkfC disk >ioscan-disk.out

check whether it shows all luns.

run syminq or inq or symdev
(all give similar outputs with slight variations). here it will show the lun ids and their paths.
From this, you can proceed with your VG configuaration.

syminq/inq will be in /opt/emc path.

HTH
raj

[Mark van Hassel]

Hi,

Check what the Target/Lun addresses are for the new LUNs. These numbers come back in the device file. i.e. T/L address 012 is device /dev/dsk/cXt1d2.

HtH,

Mark