01-10-2012 04:21 PM
At one of our sites, the disaster recovery plans for a broken EVA4400 is to throw another EVA4400 in.. connect up the disk shelf, power up.. and away she goes.
Of course, that is the plan and the second EVA4400 has already been purchased. Question are:
1. Is this possible?
2. What configuration is required to be completed on the 'standby' EVA4400 so that it will just accept the existing disks?
We are not buying storage / disks for the second EVA4400, so replication etc is not an option.
Obviously I will configure IP, hostname.. but what, if anything, is required to be configured so that all the data just comes up and working when the replacement EVA4400 is installed and powered on?
I've researched, maybe not very well, but have been unable to find any relevant information.
What I do not want to do is to hook up the disk shelf to my new EVA4400 and lose all the data.. that would be bad.. :)
Any pointers to doco, or helpful information would be greatly appreciated.
Note that I am not a storage expert by any stretch of imagination, but I did setup the existing, installed, EVA4400.. but that's about it.
01-10-2012 09:41 PM
the kind of documentation you want to find is the one that tells you what the system does when it finds a bunch of disks and there is working set of quorom disks/metadata on them.
Maybe this is only available in the EVA training materials..
I doubt what you want to do is supported.
Some questions I have (but no answers to unfortunately) to get you started:
Do you need to have the same WWN/Checksum entered on the replacement system?
Do you need to move the riser card as well?
Do the disk shelves need to have the same ID as they had on the original EVA?
Obviously firmware on the controllers and management module (and CV) should be the same versions.
Click the Kudos Star if this was a good post.
01-10-2012 09:52 PM
Your plan assumes that for whatever reasons both of the redundant controllers will completely fail with all of the disks and enclosures still working.
IMHO this is very unlikely. Based on experience you will either have a problem with a single controller or a single disk, in a worst case with a couple of disks. In such cases a complete set of new controllers will not help at all. IMHO this is a waste of money.
Hope this helps!
There are only 10 types of people in the world -
those who understand binary, and those who don't.
No support by private messages. Please ask the forum!
If you feel this was helpful please click the KUDOS! star in the left column!
01-12-2012 02:54 PM
From a technical point of view a "blank" EVA will accept your disks and boot up as long as the quorum disks are intact. Disk position is not relevant. The WWNN of the source system will also "move" to the new box since it is stored in the MLD. WWPNs will be the same as on the "source".
Assign a kudo to this post, if you find it useful.
03-20-2012 01:59 PM
We had the same thought here, but added an unlimited CA license to the mix and end up finding that another EVA with CA is an incredible solution for DR.
1) Setup the new EVA4400 on the same SAN. It doesn't really have to have the same type/size disks (as long as the overall storage volume is about the same). Mind you, don't mix FC and SATA drives.
2) Install the CA licenses on both source and destination.
3) Create the same hosts on the *new* as the *old* (saves time later on).
4) Install the MPIO bits on the hosts.
5) Replicate all the volumes on the *old* to the *new*.
If your "old" EVA ever dies, you log to the *new* EVA with Command View and reverse the replication (source<->destination), then present the volume to the proper nodes.
You might not even have to reboot the nodes, just refresh the disk management to see the volumes.
Play around with presenting the destination volume to the nodes ahead of time. I tried that and the servers instantly saw the volumes with the proper drive names as soon as i reversed the replication.
Swapping disks might work, but it's a lot less work to just have another EVA sitting there with the data already on it.