Comments on: EnableResignature and/or DisallowSnapshotLUN http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/ Building blocks for virtualization... Thu, 18 Mar 2010 02:57:03 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Hamish http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-6784 Hamish Wed, 24 Feb 2010 00:33:00 +0000 http://www.yellow-bricks.com/?p=1213#comment-6784 The misconception of LUN IDs being stored in the UUID may have something to do with the following passage from Mike Laverick's SRM book; "UUIDs are generated by using three core variables including date, time and LUN number in order to guarantee that the UUID value be absolutely unique". The misconception of LUN IDs being stored in the UUID may have something to do with the following passage from Mike Laverick’s SRM book;

“UUIDs are generated by using three core variables including date, time and LUN number in order to guarantee that the UUID value be absolutely unique”.

]]> By: Amalia http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1998 Amalia Sun, 14 Dec 2008 12:32:53 +0000 http://www.yellow-bricks.com/?p=1213#comment-1998 I have used resignaturing once when several "stand-alone" ESX servers were grouped together in a cluster. Each ESX had its own, single lun and unfortunately some where presented with identical lun numbers in the SAN config. We had to re-number some luns to be able to present them to all the ESX servers. This caused some luns to be seen as snapshot luns. I found the follwing article of much use: http://tinyurl.com/3qsldy It describes all the steps that need to be taken. I have used resignaturing once when several “stand-alone” ESX servers were grouped together in a cluster. Each ESX had its own, single lun and unfortunately some where presented with identical lun numbers in the SAN config. We had to re-number some luns to be able to present them to all the ESX servers. This caused some luns to be seen as snapshot luns.
I found the follwing article of much use: http://tinyurl.com/3qsldy

It describes all the steps that need to be taken.

]]> By: Duncan http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1996 Duncan Sat, 13 Dec 2008 21:29:11 +0000 http://www.yellow-bricks.com/?p=1213#comment-1996 Like I said, it depends. How many VC's do you guys have? If you have just 1 VC than there's no direct real need to reregister. Like I said, it depends. How many VC’s do you guys have? If you have just 1 VC than there’s no direct real need to reregister.

]]> By: happyhammer http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1995 happyhammer Sat, 13 Dec 2008 17:54:36 +0000 http://www.yellow-bricks.com/?p=1213#comment-1995 very intersting article we have come across this with a client recently using EMC recoverpoint(different SAN's, sites and clusters) and EMC best practice dictates to re signature, however if you have a 2nd VMDK on another LUN the reregistering process is very timeconsuming due to datastore having the snap-LUNNAME, however with Disallow snapshot would be far quicker as the datastore name will be the same as what teh vmx refers to If you have seperate sites, with seperate SAN's and ESX clusters then is the Disallow the best method ? very intersting article
we have come across this with a client recently using EMC recoverpoint(different SAN’s, sites and clusters) and EMC best practice dictates to re signature, however if you have a 2nd VMDK on another LUN the reregistering process is very timeconsuming due to datastore having the snap-LUNNAME, however with Disallow snapshot would be far quicker as the datastore name will be the same as what teh vmx refers to
If you have seperate sites, with seperate SAN’s and ESX clusters then is the Disallow the best method ?

]]> By: Paul O http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1991 Paul O Sat, 13 Dec 2008 09:05:37 +0000 http://www.yellow-bricks.com/?p=1213#comment-1991 I've also seen this when switching between clusters and EMC SAN's. One LUN and machines moved to a different cluster and VirtualCenter Instance. Changing the EnableResignature setting worked a treat. Also had it with some DAS and Dell servers (Direct Attached Storage) after a SCSI card firmware update and once again this setting enabled read/write on the LUN. Heart in mouth stuff certainly if you've never heard of it before. Paul I’ve also seen this when switching between clusters and EMC SAN’s. One LUN and machines moved to a different cluster and VirtualCenter Instance. Changing the EnableResignature setting worked a treat.

Also had it with some DAS and Dell servers (Direct Attached Storage) after a SCSI card firmware update and once again this setting enabled read/write on the LUN. Heart in mouth stuff certainly if you’ve never heard of it before.

Paul

]]> By: Roger Lund http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1988 Roger Lund Fri, 12 Dec 2008 21:48:02 +0000 http://www.yellow-bricks.com/?p=1213#comment-1988 I ran in to this, when I was unplugged and replugging in multiple sata drives into different ports on my unsupported ESXi box. The test box would see the drive but not the datatstore, therefore I had to enable Resignature to get it to see it again. I linked to this post on my blog. Thanks again. Roger L http://rogerlunditblog.blogspot.com/ I ran in to this, when I was unplugged and replugging in multiple sata drives into different ports on my unsupported ESXi box. The test box would see the drive but not the datatstore, therefore I had to enable Resignature to get it to see it again.

I linked to this post on my blog.

Thanks again.

Roger L

http://rogerlunditblog.blogspot.com/

]]> By: Alessandro Di Fenza http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1967 Alessandro Di Fenza Thu, 11 Dec 2008 17:26:19 +0000 http://www.yellow-bricks.com/?p=1213#comment-1967 I forgot one point: 6bis)after the rescan set EnableResignature=0 :-) I forgot one point:
6bis)after the rescan set EnableResignature=0

:-)

]]> By: Alessandro Di Fenza http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1966 Alessandro Di Fenza Thu, 11 Dec 2008 17:24:33 +0000 http://www.yellow-bricks.com/?p=1213#comment-1966 I'll test it this saturday, one of our customer will migrate his emc storage, i'll be there to monitor the vmware infrastructure. Let me explain what I think it will occur (yours comments are welcome) 1)shutdown every vm 2)shutdown every esx 3)migrate storage (using sancopy) 4)mask or remove old storage, I don't want esx to see them (point 3 and 4 will be achieved by the storage team) 5)start one esx (I hope it will mark copied lun as snapshot) 6)set EnableResignature=1 then rescan lun (I hope copied lun won't be marked as snapshot) 7)start all esx 8)reregister all vm (I think every vm will be orphaned or disconnected, old vmx point to old lun) That's all Do you think I'll be happy saturday evening? :-) I’ll test it this saturday, one of our customer will migrate his emc storage, i’ll be there to monitor the vmware infrastructure.
Let me explain what I think it will occur (yours comments are welcome)
1)shutdown every vm
2)shutdown every esx
3)migrate storage (using sancopy)
4)mask or remove old storage, I don’t want esx to see them (point 3 and 4 will be achieved by the storage team)
5)start one esx (I hope it will mark copied lun as snapshot)
6)set EnableResignature=1 then rescan lun (I hope copied lun won’t be marked as snapshot)
7)start all esx
8)reregister all vm (I think every vm will be orphaned or disconnected, old vmx point to old lun)

That’s all

Do you think I’ll be happy saturday evening? :-)

]]> By: Jason Boche http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1965 Jason Boche Thu, 11 Dec 2008 16:06:06 +0000 http://www.yellow-bricks.com/?p=1213#comment-1965 BTW, very nice post Duncan. I hope that others would verify that I know what kind of time/energy/knowledge it takes to write quality blog post like this. BTW, very nice post Duncan. I hope that others would verify that I know what kind of time/energy/knowledge it takes to write quality blog post like this.

]]> By: Jason Boche http://www.yellow-bricks.com/2008/12/11/enableresignature-andor-disallowsnapshotlun/comment-page-1/#comment-1964 Jason Boche Thu, 11 Dec 2008 16:04:01 +0000 http://www.yellow-bricks.com/?p=1213#comment-1964 I've run into one nightmare with EnableResignature and/or DisallowSnapshotLUN in my prod environment and IMO one experience with this is enough for anyone. I had not tweaked these settings previously but I was forced to get intimate with them when for some strange reason, clustered ESX hosts started seeing existing SAN LUNs with new names, basically the same correct name, but with the infamous (1) behind it. So basically HOSTA sees a LUN as LUN1 and HOSTB sees the same LUN as LUN1(1). Getting things settled down was a pain in the ass and I was sweating it for a little while. I’ve run into one nightmare with EnableResignature and/or DisallowSnapshotLUN in my prod environment and IMO one experience with this is enough for anyone.

I had not tweaked these settings previously but I was forced to get intimate with them when for some strange reason, clustered ESX hosts started seeing existing SAN LUNs with new names, basically the same correct name, but with the infamous (1) behind it. So basically HOSTA sees a LUN as LUN1 and HOSTB sees the same LUN as LUN1(1).

Getting things settled down was a pain in the ass and I was sweating it for a little while.

]]>