asked on
EMC Clariion/VNX MirrorView - testing secondary images
I had this question after viewing EMC Clariion MirrorView - testing secondary images.
The previous message dealt with Clariion disk arrays. I have a very similar configuration, with two VNX 5200 arrays. I have a MirrorView connection between them, and am testing with a 50 GB LUN on the primary system. That LUN is mirrored to a secondary array, using Mirrorview/A.
Just as with the previous question, my goal is to test the secondary (remote) image without impacting the primary image, specifically without promoting it.
My scenario: Primary Data center array has the 50 GB LUN in it. Since I'm using MirrorView/A, it creates a mirror on the primary array, and a remote image on the secondary array. No problem here. The primary and secondary are fully synchronized.
As Duncan suggested in the related post, I created a clone (of the remote image) on the secondary array. I then let the clone synchronize, and waited until it was at 100%. All good so far. I then fractured the clone, and added the clone to a storage group on my remote ESX hosts (6.0, in case it makes any difference.)
I haven't turned the re-signaturing of the VMFS partition on, which may be part of my problem. What happens is that yes, the ESX hosts see the LUN, but when I try to add it, I get an error message that reads "Cannot change the host configuration."
Previous to this attempt, I had created a similar 50 GB LUN on my primary array, and mirrored that (also MirrorView/A) to the secondary array. In that attempt, I fractured the MirrorView connection, added the remote image to a storage group, and tried adding the remote LUN as a datastore to the remote ESX host. I got the same error message - "Cannot change the host configuration." I opened a case with VMware on that, and they confirmed that the host sees the new LUN, but doesn't see any partitions on it.
I thought the problem might be that I hadn't promoted the secondary image to be primary, but in my case that's exactly what I want to avoid - I want the primary image to stay in a production status, and do my D/R testing on the secondary array.
I'll try issuing the esxcfg-advcfg -s 1 /LVM/EnableResignature command on the host tomorrow, to see if that makes any difference. Beyond that, does anyone have any advice on how to make this work? EMC's advice seems to consistently be to bring the production servers down, synchronize the mirrors, promote the secondary LUN, do the D/R testing, shut them down, and then promote the primary. Unfortunately, for large LUNs (think upwards of 5 TB), that's not a workable solution, as the time to re-synchronize with the primary LUNs is going to be far too long to do this in any kind of a test window.
Any insights or advice on this would be greatly appreciated!
The previous message dealt with Clariion disk arrays. I have a very similar configuration, with two VNX 5200 arrays. I have a MirrorView connection between them, and am testing with a 50 GB LUN on the primary system. That LUN is mirrored to a secondary array, using Mirrorview/A.
Just as with the previous question, my goal is to test the secondary (remote) image without impacting the primary image, specifically without promoting it.
My scenario: Primary Data center array has the 50 GB LUN in it. Since I'm using MirrorView/A, it creates a mirror on the primary array, and a remote image on the secondary array. No problem here. The primary and secondary are fully synchronized.
As Duncan suggested in the related post, I created a clone (of the remote image) on the secondary array. I then let the clone synchronize, and waited until it was at 100%. All good so far. I then fractured the clone, and added the clone to a storage group on my remote ESX hosts (6.0, in case it makes any difference.)
I haven't turned the re-signaturing of the VMFS partition on, which may be part of my problem. What happens is that yes, the ESX hosts see the LUN, but when I try to add it, I get an error message that reads "Cannot change the host configuration."
Previous to this attempt, I had created a similar 50 GB LUN on my primary array, and mirrored that (also MirrorView/A) to the secondary array. In that attempt, I fractured the MirrorView connection, added the remote image to a storage group, and tried adding the remote LUN as a datastore to the remote ESX host. I got the same error message - "Cannot change the host configuration." I opened a case with VMware on that, and they confirmed that the host sees the new LUN, but doesn't see any partitions on it.
I thought the problem might be that I hadn't promoted the secondary image to be primary, but in my case that's exactly what I want to avoid - I want the primary image to stay in a production status, and do my D/R testing on the secondary array.
I'll try issuing the esxcfg-advcfg -s 1 /LVM/EnableResignature command on the host tomorrow, to see if that makes any difference. Beyond that, does anyone have any advice on how to make this work? EMC's advice seems to consistently be to bring the production servers down, synchronize the mirrors, promote the secondary LUN, do the D/R testing, shut them down, and then promote the primary. Unfortunately, for large LUNs (think upwards of 5 TB), that's not a workable solution, as the time to re-synchronize with the primary LUNs is going to be far too long to do this in any kind of a test window.
Any insights or advice on this would be greatly appreciated!
* emcStorageVMwareDisaster Recovery
Last Comment
Andrew Hancock (VMware vExpert PRO / EE Fellow/British Beekeeper)