Hi,
Is it possible to use an EMC Clariion CX400 with ESX 4.0? HCL says no - but are there any tricks?
Regards,
Mike
Old EMC product are not (yet) in new HCL.
But this does not mean that they do no work
I used a CX 300 and CX 500 without problem and without tricks.
Andre
Ahhh - great news! So my yesterday's potential bargain on eBay at least eventually wasn't complete bunk (766 Euroes for CX400 with 1xDAE (15x73GB FC), 1xDPE (15x73GB FC) and 1xUPS (new batteries) - including all cables and software).
In my amateurish opinion the CX300, CX500 and CX400 should be hardware compatible. So if yours worked with ESX 4, mine should work also (I hope).
What kind of HBA did you use? QLogic I suppose. I'm currently trying to shoot a QLE2462 or QLE2464...would you think that would do? Also I don't own a SAN-Switch...but my goal is anyway only to use the CX400 directly attached to the ESX host. Is that possible?
What kind of HBA did you use? QLogic I suppose
On this environment I've old Emulex LP1050 and Qlogic QLA2432
Andre
Do you have the storage directly attached or connected it over a SAN-Switch?
I could currently get a Brocade Silkworm 3800...would that do it?
I use FC switches.
Brocade could be fine.
Andre
Hi Andre,
I've finally got my hands on a pair of QLA2342s...ESX correctly recognizes both ports as vmhba2 and vmhba3. I've connected the HBA via Fibre directly to Port-0 on the CX400. But although I did completely reconfigure the both enclosures from scratch (via the Navisphere browser plugin) the resulting LUNs won't show up in ESX when trying to add storage. Do you have a clue where my problem comes from?
Do I have to do something from the service console? Some people talk about "esxcfg-vmhbadevs", a command that doesn't seems to be available on my ESX 4.
Where could I start?
Best Regards,
Mike
Hi Andre,
I've finally got my hands on a pair of QLA2342s...ESX correctly recognizes both ports as vmhba2 and vmhba3. But although I did completely reconfigure the both enclosures from scratch (via the Navisphere browser plugin) the resulting LUNs won't show up in ESX when trying to add storage. Do you have a clue where my problem comes from?
Do I have to do something from the service console? Some people talk about "esxcfg-vmhbadevs", a command that doesn't seems to be available on my ESX 4.
Where could I start?
Best Regards,
Mike
If you see the HBA in ESX you have finished on this side
Use Navishpere to see if you have to host listed.
Check in your Storage Group: you must have all your ESX hosts and all the VMFS(+RDM if you have) LUNs.
Andre
Hi - thanks for tha fast feedback!
My hardware is a CX400 with one DPE2 and one DAE2 and the UPS device. I also have single ESX 4.0-164009 which is working perfectly with Vmware Client or via a Vcenter server. In Navisphere: I created 2 RAID-5 arrays of 12 disks each, for each 2 others are Hotspare. I bound them to LUN0 und LUN1. I created a Storage Group 1 and added the LUNs there. I went to Connection Status and added the ESX server's HBA's WWN (selecting as host my ESX server (which is found and added to the Storage Group correctly), and also ArrayCommPath=Enabled, Initiator Type Clariion Open, FailOver Mode 0 and the Storage Group from before, Unit SerialNumer=LUN). I selected SP A 0/1 and SP b 0/1 and the optical cable is attached directly from the ESX server's HBA (Port-0) to the leftmost LC port in the Clariion.
Reboting the ESX or rescanning for new devices doesn't change anything; nothing found. Also there are no devices found in the QLogic's firmware.
Do you have something for me - am I missing something or looking toooo dumb at it?
Regards
Mike
I created 2 RAID-5 arrays of 12 disks each, for each 2 others are Hotspare.
I suggest to use max 7-9 disks in a RAID5 RAID Group on EMC Clariion product.
Also the first 5 disks contain the EMC OS partition and are "smaller" than the other, so could be useful to have in separated RAID Group.
With 2 DAE, you can also use 1 global hostspare (but must the the bigger disk).
How big are your LUN? Are smaller than 2 TB?
Andre
Are you sure your esx servers are not still listed in navishere using the old wwn's? I never used powerpath before so i would have to manually create the hosts with the wwn's in naviphere (and make sure the zones were correct in the FC switches). When switching fc cards I woulf remove that registration and re-add the server with the correct wwn's
In esx4 you will still need software ot have the autoregistration happen in navisphere
One confusing thing - I selected SP A 0/1 and
SP b 0/1 and the optical cable is attached directly from the ESX
server's HBA (Port-0) to the leftmost LC port in the Clariion.
Didn't you say you had FC switches? why are you plugged right into the SP''s?
I can build new RAID Groups when the basic problems are solved.
As long as they aren't the number of disks in them doesn't matter.
As I wrote in my last posting, I don't have two DAEs but only one. The other device is the DPE. The disks are 73Gbytes each, so no LUNs bigger that 2TByte.
But what about the basic problems of no LUNs being found a) by the QLogic HBA and b) by ESX?
No I don't have a switch (yet), the HBA is directly connected with the CX400's DPE. Do I need a switch?
What old WWNs? I have one new HBA (shot on 'bay last week) and there never was another one registered to the CX400 since I own it.
As I told before I manually registered the HBA's ports in Navisphere and bound them to the ESX Host. Never the less the LUNs don't show up there. What software would I need on the the ESX' side? The NaviAgent is an optional piece of software - the LUNs should show with or without it. Btw. EMC now takes the third week of clearing my account for the powerbla-portal, so I couldn't even download that software if I needed it.
Do I need a switch?
No, also a DAS solution like yours must work.
But you must be sure that each host WWN is connected and logged in your connection manager.
And that are correctly associated to the host name/IP.
Andre
I've set them up in connection manager, but their icons are greyed out and they are not logged in. I have one adapter with two ports, each port has an own WWN; each port shows up in ESX as seperate vmhba. On the EMC Port 0 of SP A is connected directly to Port 0 of the HBA. Both LUNs are bound to SPA-0. In connection manager I've bound the WWNs of both ports to SPA-0. They register, but are not "Fibre Logged In".
You have also to have in login. Could be a link problem. Your CX400 is FC2, maybe your HBA are FC4?
Use Group Edit to build to group the HBA to the right host.
Andre
Yes I know; not being "Fibre Logged in" is definitely my central problem. The QLA2342 is FC2 too. It has the latest Firmware (1.54) and is set to Point-to-Point-Connection and 2Gbit (not AutoSelect). Group Edit does nothing more I have already done - register a HBA in conjunction with a host. Group Edit just simplyfies similar operations on several HBAs. What do you set as values for "ArrayCommPath" and "UnitSerialNumber"?
I always use the default values (when I use manual registration):
"ArrayCommPath" -> enabled
"UnitSerialNumber" -> Array
But I prefer use naviagent to do host registration.
Andre
I would very like to use the Agent, but due to EMC's strange information policy I am unable to get full Powerlink access. Do you have? Could you download the latest packages and upload them somewhere for me?