Relocation of DAE

Question

I tried alot to find any kind of documentation but failed, so here I am and would appreciate any kind of help.

I need to move one of the DAEs from one clariion (CX4-120) and install it on another (CX4-120).

The DAE has no data/luns/raid group on it.

We have a total of 3 DAEs on both the clariions and the one to be removed is Bus 0 Enclosure 1.

I would guess that its better to have EMC do this kind of work but

1. Is there any way that we do it ourselves, if yes how.

2. The DAE which is stack cabled on top of the DAE mentioned above, will the hosts connected to it lose connectivity.

Again, any kind of input will be greatly appreciated.

Sri

kelleg · Answer

It would be best to have support do this as it is a configuration change and they need to be aware of that.

Do you have three DAE's on each CX4-120? 0-0, 0-1, and 0-2? If you remove 0-1, you will disconnect 0-2 which is the next in line. When you disconnect the cables, that will break connection to the disks. If you only remove one cable, let the hosts failover the LUNs to the other SP, then inset the cable to the last DAE, then do the same on the other side, you should be able to remove the DAE without needing to take down the whole array.

When you add the DAE to the other array, you'll need to be careful about what the enclosure numbering will be.

glen

Sridhar246 · Answer

Yes we have three DAEs on each Cx4-120...0-0, 0-1, 0-2 and 0-1 will be moved over to other Cx4-120 as 0-3. Other than disconnecting one LCC on each DAE at a time do we have to do anything more?

kenn2347 · Answer

One thing i would recommend is that if your going to do this yourself, that you use the USM. it should step you through the process. The USM might not do this for you as there are some steps that should be completed by EMC or a Certified Partner and those steps are found in the Proceedure Generator.

Are the disks in 0_2 the same size and type as the disks in the DAE you are removing? IF so, you can create the same raid groups and luns on 0_1 that are on 0_2 and then do a Lun Migration to move all the data off 0_2 and onto 0_1. Once that is done, you can then remove 0_2 and you wont have a gaps in your enclosure ID loop. You could then use the USM to remove that DAE since doing it that way does not require any exrta expertise.

Those lun migrations are done on the SAN and has no impact on your host.

kelleg · Answer

Just make sure your host failover is working correctly - the LUNs in 0-2 will get trespassed and one LCC cable is moved up.

I like Kenn's suggestion better - not leaving a gap in the DAE stacking and not having a gap in the addressing on the DAE's - nice suggestion Kenn - also you can just power off the DAE when the migrations are finished and remove the cables without worrying about the hosts.

glen

kelleg · Answer

USM is the Unisphere Service Manager - from PowerLink go to:

Home > Support > Software Downloads and Licensing > Downloads T-Z > Unisphere Service Manager (USM)

You'll also need the latest version of Navisphere CLI

Home > Support > Software Downloads and Licensing > Downloads J-O > Navisphere Server Software

glen

Sridhar246 · Answer

No the disks on 0_1 and 0_2 are not of the same size...I can still rearrange the storage around but I will end up putting some luns on the vault disks. These luns (ones which will end up on vault disks) don't get alot of IOPs but I am not sure of the exact numbers.

Also is USM Unispere Storage Manager?

kenn2347 · Answer

And i would recommend upgrading to the current version of FLARE.  there have been a bunch of improvments since Release 28

Vipin VK · Answer

Hi Sridhar,          You can use USM with this FLARE code also.           Vipin V.K

Sridhar246 · Answer

The Cx4-120 is running on flare code 4-28 and we are not using Unisphere to manage it...can i still use USM?

christopher_ime · Answer

>> Just make sure your host failover is working correctly - the LUNs in 0-2 will get trespassed and one LCC cable is moved up. Simply wanted to comment that if running at least FLARE 26, the lower-redirector will route I/O through the peer SP so a trespass should not occur in the case of back-end 'failures' such as the temporary loss of connectivity to each LCC as proposed in the procedure above. Despite sounding like an ALUA feature (which is implemented in the upper-redirector), the behavior with the lower-redirector benefits all hosts and is entirely independent of host configuration (such as fail-over mode and host-based failover software). This is described in the: EMC CLARiiON Asymmetric Active/Active Feature whitepaper available on PowerLink.

Sridhar246 · Answer

Thanks all for your replies...we ended up bringing the whole clariion down (we were told by EMC that this is the only way...dont actually need EMC if we are bringing the whole environment down but thats ok...) and removing the cabling at one site and recabling at other.

kelleg · Answer

Please remember to mark this thread as Answered if your qusetion was answered. glen

CLARiiON

Relocation of DAE

Was this post helpful?