SATA Drives and RDM

[ibforbes]

I know I should have checked first but it never crossed my mind in the scramble.

Last night we added a tray of SATA drives to our EMC Clarion. The intent was to connect a couple large RDMs to an existing VM that hosts a great deal of static data. Everything was normal right up until I went to add and RDM.

Apparently when a LUN created on the SATA drives is presented to the ESX hosts the RDM option is greyed out. I was able to create VMFS partitions on these luns instead and add the disk that way.

Here's my concern.... I have a number of VMs that currently have RDMs attached. They continue to work but new RDMs are not possible. What is the workaround or best solution for managing this. Should I break off part of my farm to host only machines with RDMs and disconnect the SATA drives from those hosts or is there some setting in ESX that will make this all work?

I'm running vCenter 2.5 update 4 and ESX 3.5 update 2.

Thanks,

Bruce

[AndreTheGiant]

To add a new RDM disk you have to follow this step:

• build a LUN (smaller than 2TB)

• add in the same Storage Group of all your ESX

• rescan on each ESX (under storage adapter): you must see the new LUN, but do not add new datastore

• on VM, add a new RDM disk

Andre

[ibforbes]

Thanks. Actually I am VERY familiar with how to add RDMs. The problem

is how to add them now that the SATA drives are in place? As it turns,

now that these SATA drives have been added, I actually have to

completely reboot all of the servers after creating the new LUN. Once I

do that the RDM appears again. If I just rescan (no matter how many

times I rescan luns) then the RDM option remains greyed out. Oddly,

that same LUN that cannot be made into an RDM is seen after a rescan and

can be made into a VMFS partition..... just not available as an RDM?

As you can imagine this is not an ideal situation if I have to reboot

every ESX host each time I want to add an RDM. This behaviour did not

exist prior to the addition of these SATA drives. Support didn't have

an answer for me on this one. Once we found they showed up after a

reboot that became the fix. The suggestion is of course upgrade to

update 4 and see what happens. Unfortunately we can't do that since the

DR side has to be upgraded first and I have not been successful in

getting the people that manage the DR servers to do the upgrade.

[AndreTheGiant]

It does not depend by disks technology.

On a storage LUNs are LUNs (not at the same performance ).

I've got a similar problem on a ESX 3.5 (U1 o U2) with 4 ESXs in a cluster and a lot of LUNs.

I've resolved it upgrading to last version (U4), removing all RDMs, rebooting the ESXs and adding RDM one by one to all ESXs.

Andre

[ibforbes]

Interesting.... in theory you are correct but not entirely..... take a

look if you have SATA drives and you will find they are clearly

identified by ESX as SATA drives so the OS does know. I never had any

issues adding RDMs through the normal routine until I added these SATA

drives.

I'm in process of getting the DR site upgraded to U4 so I can upgrade

the primary site and then we'll see what happens.

[AndreTheGiant]

take a look if you have SATA drives and you will find they are clearly identified by ESX as SATA drives so the OS does know.

This is my identifier:

Vendor: DGC Model: RAID 5

No info on type of disks.

Which version of Flare do you have? Have you called EMC?

Andre

[ibforbes]

Flare 03.26.020.5.025 on a Clarion CX3-20.

I just added another experimental LUN and you ar correct. It was in

Navisphere that I see it identified as SATA not in ESX.... although with

this same version of ESX I never had this issue until the SATA drives

were introduced. Coincedence perhaps?

[AndreTheGiant]

Coincedence perhaps?

Yes, or some storage problem...

Each SATA LUN is smaller than 2 TB?

Andre

[ibforbes]

Yes... the largest one is 1.3TB.

[AndreTheGiant]

On Navisphere do you have some errors or warning?

Andre

[ibforbes]

No errors at all in Navisphere. Everything is normal on both sides.

EMC installed the drives and were comfortable when they left as well so

I have to assume no issues there. VMware support feels the issue is in

U2 and installing U4 will likely solve it. The word is from the people

that manage the DR farm is they think they MIGHT be able to get this

done by the end of August (4 servers) so my 12 servers that I can easily

do in a day will have to wait until then.

PLEASE BE ADVISED:

The information transmitted herein is intended only for the named recipient(s) above and may contain information that is privileged, confidential or exempt from disclosure under applicable law. If you have received this message in error, or are not the named recipient(s), please immediately notify the sender toll free 1-800-668-6203 (in Canada or US) or by reply to this communication and delete this message from any computer. Thank you.

Please consider the environment before printing this email.