Correct. Same filer. Thanks
----- Original Message ----- From: Milazzo Giacomo G.Milazzo@sinergy.it To: Patrick van Helden pvh@databasement.eu; Klise, Steve; toasters@mathworks.com toasters@mathworks.com Sent: Fri May 16 10:29:55 2008 Subject: R: MSCS and snapmirror question
Of course but that guy said he's migrating on the same filer to another aggregate.
________________________________
Da: Patrick van Helden [mailto:pvh@databasement.eu] Inviato: ven 16/05/2008 19.12 A: Milazzo Giacomo; Klise, Steve; toasters@mathworks.com Oggetto: RE: MSCS and snapmirror question
Guys,
Please be aware that MSCS will keep track of the lun serials. So you would probably need to set the lun serial numbers back to the original ones when your snapmirroring to another filer.
Regards,
Patrick van Helden Databasement BV pvh@databasement.eu
-----Oorspronkelijk bericht----- Van: owner-toasters@mathworks.com namens Milazzo Giacomo Verzonden: vr 5/16/2008 18:01 Aan: Klise, Steve; toasters@mathworks.com Onderwerp: R: MSCS and snapmirror question
You will not have issues. And you don't need to update your mirror. If you simply power off you clusters no i/o activity will be performer on LUNs. Then you can simply copy (vol copy) the volume(s) containing your LUNs, rename the new volumes as the old (in case of iSCSI LUNs and if you don't want to recreate the connections) and restart you clusters.
In case your nodes (but it will not happen) will loose the LUNs with Snapdrive you can safely reconnect them starting with the quorum one, on first node then restarting the cluster service, then on the second node and so on with all other cluster shared LUNs
Regards
-----Messaggio originale----- Da: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] Per conto di Klise, Steve Inviato: venerdì 16 maggio 2008 15.50 A: toasters@mathworks.com Oggetto: MSCS and snapmirror question
I am migrating from an aggregate to another aggregate in the same filer (6030). The LUNS are attached to print clusters, 2 node and 3 node clusters (WIndows 2003 sp2 x86 etc).
I have snapmirrored the volumes. If I stop the cluster service on the windows boxes, do a snapmirror update, the disconnected the quorum and other resource disks, do a snapmirror update, release the snapmirror, remove and replace the disks with snapdrive with the same disk letters. fire up the cluster service.. Will that work? has anyone done that before? Any gotchas?
Any help is greatly appreciate!
Steve,
As long as you keep the resources offline and bring them back up with the same drive letter, you are generally OK. It's when you try changing drive letters and assume setting the disk dependency will work is how you run into trouble. I like doing the procedure you described but we avoid snapmirring the Quorum (it will work, but over here we just don't like to offline the cluster resource)
With the Quorum / MSDTC LUN, it is pretty painless to do a non-intrusive migration of the data - we usually end up doing it twice because we want to leave Q: as the quorum drive letter:
Add temp quorum as W: Open MSDTC component services panel, offline msdtc Change quorum drive from q: to w: (right click on cluster, go to quorum tab, choose W: drive) Change msdtc log path to w:\ in MSDTC component services control panel Online msdtc, test failover. If failover is successful, disconnect legacy quorum Add new final quorum as Q: Repeat process to move to Q:\ Test cluster failover
Patrick,
In regards to the LUN serials changing, I've never seen this to be an issue with MSCS, but I've heard it can be a big gotcha when snapmirroring VMWare LUNs. Opinions?
- Hadrian
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Klise, Steve Sent: Friday, May 16, 2008 10:50 AM To: G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: Re: R: MSCS and snapmirror question
Correct. Same filer. Thanks
----- Original Message ----- From: Milazzo Giacomo G.Milazzo@sinergy.it To: Patrick van Helden pvh@databasement.eu; Klise, Steve; toasters@mathworks.com toasters@mathworks.com Sent: Fri May 16 10:29:55 2008 Subject: R: MSCS and snapmirror question
Of course but that guy said he's migrating on the same filer to another aggregate.
________________________________
Da: Patrick van Helden [mailto:pvh@databasement.eu] Inviato: ven 16/05/2008 19.12 A: Milazzo Giacomo; Klise, Steve; toasters@mathworks.com Oggetto: RE: MSCS and snapmirror question
Guys,
Please be aware that MSCS will keep track of the lun serials. So you would probably need to set the lun serial numbers back to the original ones when your snapmirroring to another filer.
Regards,
Patrick van Helden Databasement BV pvh@databasement.eu
-----Oorspronkelijk bericht----- Van: owner-toasters@mathworks.com namens Milazzo Giacomo Verzonden: vr 5/16/2008 18:01 Aan: Klise, Steve; toasters@mathworks.com Onderwerp: R: MSCS and snapmirror question
You will not have issues. And you don't need to update your mirror. If you simply power off you clusters no i/o activity will be performer on LUNs. Then you can simply copy (vol copy) the volume(s) containing your LUNs, rename the new volumes as the old (in case of iSCSI LUNs and if you don't want to recreate the connections) and restart you clusters.
In case your nodes (but it will not happen) will loose the LUNs with Snapdrive you can safely reconnect them starting with the quorum one, on first node then restarting the cluster service, then on the second node and so on with all other cluster shared LUNs
Regards
-----Messaggio originale----- Da: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] Per conto di Klise, Steve Inviato: venerdì 16 maggio 2008 15.50 A: toasters@mathworks.com Oggetto: MSCS and snapmirror question
I am migrating from an aggregate to another aggregate in the same filer (6030). The LUNS are attached to print clusters, 2 node and 3 node clusters (WIndows 2003 sp2 x86 etc).
I have snapmirrored the volumes. If I stop the cluster service on the windows boxes, do a snapmirror update, the disconnected the quorum and other resource disks, do a snapmirror update, release the snapmirror, remove and replace the disks with snapdrive with the same disk letters. fire up the cluster service.. Will that work? has anyone done that before? Any gotchas?
Any help is greatly appreciate!
This worked like a charm over the weekend!
Interesting note. I had a 3 node cluster, and when I would try and change the disk letter on a node that did not own the resource, it failed. Note to self. On a multi-node cluster, have all the resources from the machine you are changing the drive letters.
Turns out you can change the quorum resource pretty easy. You right click on the cluster, properties, cluster tab, change it from drive letter to drive letter, done. I just deleted and recreated the msdtc resource.
Done
Another question:
I have to migrate a bunch of sql clusters from a 940/960's to 3020's 6030's. We use Snap Drive for SQL. Should I just use SMS to migrate? One database is pretty big; Its like 200GB, and I prefer the least amount of downtime if possible. Yes, I would imagine I would have to set the LUN serial number, not sure how to do that.
Thanks everyone.
-----Original Message----- From: Hadrian Baron [mailto:Hadrian.Baron@vegas.com] Sent: Monday, May 19, 2008 8:58 AM To: Klise, Steve; G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: RE: R: MSCS and snapmirror question
Steve,
As long as you keep the resources offline and bring them back up with the same drive letter, you are generally OK. It's when you try changing drive letters and assume setting the disk dependency will work is how you run into trouble. I like doing the procedure you described but we avoid snapmirring the Quorum (it will work, but over here we just don't like to offline the cluster resource)
With the Quorum / MSDTC LUN, it is pretty painless to do a non-intrusive migration of the data - we usually end up doing it twice because we want to leave Q: as the quorum drive letter:
Add temp quorum as W: Open MSDTC component services panel, offline msdtc Change quorum drive from q: to w: (right click on cluster, go to quorum tab, choose W: drive) Change msdtc log path to w:\ in MSDTC component services control panel Online msdtc, test failover. If failover is successful, disconnect legacy quorum Add new final quorum as Q: Repeat process to move to Q:\ Test cluster failover
Patrick,
In regards to the LUN serials changing, I've never seen this to be an issue with MSCS, but I've heard it can be a big gotcha when snapmirroring VMWare LUNs. Opinions?
- Hadrian
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Klise, Steve Sent: Friday, May 16, 2008 10:50 AM To: G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: Re: R: MSCS and snapmirror question
Correct. Same filer. Thanks
----- Original Message ----- From: Milazzo Giacomo G.Milazzo@sinergy.it To: Patrick van Helden pvh@databasement.eu; Klise, Steve; toasters@mathworks.com toasters@mathworks.com Sent: Fri May 16 10:29:55 2008 Subject: R: MSCS and snapmirror question
Of course but that guy said he's migrating on the same filer to another aggregate.
________________________________
Da: Patrick van Helden [mailto:pvh@databasement.eu] Inviato: ven 16/05/2008 19.12 A: Milazzo Giacomo; Klise, Steve; toasters@mathworks.com Oggetto: RE: MSCS and snapmirror question
Guys,
Please be aware that MSCS will keep track of the lun serials. So you would probably need to set the lun serial numbers back to the original ones when your snapmirroring to another filer.
Regards,
Patrick van Helden Databasement BV pvh@databasement.eu
-----Oorspronkelijk bericht----- Van: owner-toasters@mathworks.com namens Milazzo Giacomo Verzonden: vr 5/16/2008 18:01 Aan: Klise, Steve; toasters@mathworks.com Onderwerp: R: MSCS and snapmirror question
You will not have issues. And you don't need to update your mirror. If you simply power off you clusters no i/o activity will be performer on LUNs. Then you can simply copy (vol copy) the volume(s) containing your LUNs, rename the new volumes as the old (in case of iSCSI LUNs and if you don't want to recreate the connections) and restart you clusters.
In case your nodes (but it will not happen) will loose the LUNs with Snapdrive you can safely reconnect them starting with the quorum one, on first node then restarting the cluster service, then on the second node and so on with all other cluster shared LUNs
Regards
-----Messaggio originale----- Da: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] Per conto di Klise, Steve Inviato: venerdì 16 maggio 2008 15.50 A: toasters@mathworks.com Oggetto: MSCS and snapmirror question
I am migrating from an aggregate to another aggregate in the same filer (6030). The LUNS are attached to print clusters, 2 node and 3 node clusters (WIndows 2003 sp2 x86 etc).
I have snapmirrored the volumes. If I stop the cluster service on the windows boxes, do a snapmirror update, the disconnected the quorum and other resource disks, do a snapmirror update, release the snapmirror, remove and replace the disks with snapdrive with the same disk letters. fire up the cluster service.. Will that work? has anyone done that before? Any gotchas?
Any help is greatly appreciate!
Steve,
SMSQL will take longer to migrate the data than Snapmirror. SMSQL is really useful when migrating the system databases, but if you use the same general approach as last time - Offline the SQL resources, remove disk dependancies, snapmirror update / quiesce / break, detach old luns, attach new luns with same drive letter, online SQL, you can be back up in minutes once you get the process down pat.
SMSQL will do a manual copy of the data from one LUN to another. I don't believe SMSQL is Snapmirror-aware enough to make this faster than traditional Snapmirror.
- Hadrian
-----Original Message----- From: Klise, Steve [mailto:klises@caminomedical.org] Sent: Monday, May 19, 2008 9:26 AM To: Hadrian Baron; G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: RE: R: MSCS and snapmirror question
This worked like a charm over the weekend!
Interesting note. I had a 3 node cluster, and when I would try and change the disk letter on a node that did not own the resource, it failed. Note to self. On a multi-node cluster, have all the resources from the machine you are changing the drive letters.
Turns out you can change the quorum resource pretty easy. You right click on the cluster, properties, cluster tab, change it from drive letter to drive letter, done. I just deleted and recreated the msdtc resource.
Done
Another question:
I have to migrate a bunch of sql clusters from a 940/960's to 3020's 6030's. We use Snap Drive for SQL. Should I just use SMS to migrate? One database is pretty big; Its like 200GB, and I prefer the least amount of downtime if possible. Yes, I would imagine I would have to set the LUN serial number, not sure how to do that.
Thanks everyone.
-----Original Message----- From: Hadrian Baron [mailto:Hadrian.Baron@vegas.com] Sent: Monday, May 19, 2008 8:58 AM To: Klise, Steve; G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: RE: R: MSCS and snapmirror question
Steve,
As long as you keep the resources offline and bring them back up with the same drive letter, you are generally OK. It's when you try changing drive letters and assume setting the disk dependency will work is how you run into trouble. I like doing the procedure you described but we avoid snapmirring the Quorum (it will work, but over here we just don't like to offline the cluster resource)
With the Quorum / MSDTC LUN, it is pretty painless to do a non-intrusive migration of the data - we usually end up doing it twice because we want to leave Q: as the quorum drive letter:
Add temp quorum as W: Open MSDTC component services panel, offline msdtc Change quorum drive from q: to w: (right click on cluster, go to quorum tab, choose W: drive) Change msdtc log path to w:\ in MSDTC component services control panel Online msdtc, test failover. If failover is successful, disconnect legacy quorum Add new final quorum as Q: Repeat process to move to Q:\ Test cluster failover
Patrick,
In regards to the LUN serials changing, I've never seen this to be an issue with MSCS, but I've heard it can be a big gotcha when snapmirroring VMWare LUNs. Opinions?
- Hadrian
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Klise, Steve Sent: Friday, May 16, 2008 10:50 AM To: G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: Re: R: MSCS and snapmirror question
Correct. Same filer. Thanks
----- Original Message ----- From: Milazzo Giacomo G.Milazzo@sinergy.it To: Patrick van Helden pvh@databasement.eu; Klise, Steve; toasters@mathworks.com toasters@mathworks.com Sent: Fri May 16 10:29:55 2008 Subject: R: MSCS and snapmirror question
Of course but that guy said he's migrating on the same filer to another aggregate.
________________________________
Da: Patrick van Helden [mailto:pvh@databasement.eu] Inviato: ven 16/05/2008 19.12 A: Milazzo Giacomo; Klise, Steve; toasters@mathworks.com Oggetto: RE: MSCS and snapmirror question
Guys,
Please be aware that MSCS will keep track of the lun serials. So you would probably need to set the lun serial numbers back to the original ones when your snapmirroring to another filer.
Regards,
Patrick van Helden Databasement BV pvh@databasement.eu
-----Oorspronkelijk bericht----- Van: owner-toasters@mathworks.com namens Milazzo Giacomo Verzonden: vr 5/16/2008 18:01 Aan: Klise, Steve; toasters@mathworks.com Onderwerp: R: MSCS and snapmirror question
You will not have issues. And you don't need to update your mirror. If you simply power off you clusters no i/o activity will be performer on LUNs. Then you can simply copy (vol copy) the volume(s) containing your LUNs, rename the new volumes as the old (in case of iSCSI LUNs and if you don't want to recreate the connections) and restart you clusters.
In case your nodes (but it will not happen) will loose the LUNs with Snapdrive you can safely reconnect them starting with the quorum one, on first node then restarting the cluster service, then on the second node and so on with all other cluster shared LUNs
Regards
-----Messaggio originale----- Da: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] Per conto di Klise, Steve Inviato: venerdì 16 maggio 2008 15.50 A: toasters@mathworks.com Oggetto: MSCS and snapmirror question
I am migrating from an aggregate to another aggregate in the same filer (6030). The LUNS are attached to print clusters, 2 node and 3 node clusters (WIndows 2003 sp2 x86 etc).
I have snapmirrored the volumes. If I stop the cluster service on the windows boxes, do a snapmirror update, the disconnected the quorum and other resource disks, do a snapmirror update, release the snapmirror, remove and replace the disks with snapdrive with the same disk letters. fire up the cluster service.. Will that work? has anyone done that before? Any gotchas?
Any help is greatly appreciate!
Eheheh :-) There's somebody needs to study better and deeper MSCS...That's the way it works as from design :-) Don't forget that also if Snapdrive make very easy to create from quorum to other disk resource MSCS is always Microsoft software :-)
For which concerns SQL migration I completely agree with Hadrian's instructions.
-----Messaggio originale----- Da: Klise, Steve [mailto:klises@caminomedical.org] Inviato: lunedì 19 maggio 2008 18.26 A: Hadrian Baron; Milazzo Giacomo; pvh@databasement.eu; toasters@mathworks.com Oggetto: RE: R: MSCS and snapmirror question
This worked like a charm over the weekend!
Interesting note. I had a 3 node cluster, and when I would try and change the disk letter on a node that did not own the resource, it failed. Note to self. On a multi-node cluster, have all the resources from the machine you are changing the drive letters.
Turns out you can change the quorum resource pretty easy. You right click on the cluster, properties, cluster tab, change it from drive letter to drive letter, done. I just deleted and recreated the msdtc resource.
Done
Another question:
I have to migrate a bunch of sql clusters from a 940/960's to 3020's 6030's. We use Snap Drive for SQL. Should I just use SMS to migrate? One database is pretty big; Its like 200GB, and I prefer the least amount of downtime if possible. Yes, I would imagine I would have to set the LUN serial number, not sure how to do that.
Thanks everyone.
-----Original Message----- From: Hadrian Baron [mailto:Hadrian.Baron@vegas.com] Sent: Monday, May 19, 2008 8:58 AM To: Klise, Steve; G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: RE: R: MSCS and snapmirror question
Steve,
As long as you keep the resources offline and bring them back up with the same drive letter, you are generally OK. It's when you try changing drive letters and assume setting the disk dependency will work is how you run into trouble. I like doing the procedure you described but we avoid snapmirring the Quorum (it will work, but over here we just don't like to offline the cluster resource)
With the Quorum / MSDTC LUN, it is pretty painless to do a non-intrusive migration of the data - we usually end up doing it twice because we want to leave Q: as the quorum drive letter:
Add temp quorum as W: Open MSDTC component services panel, offline msdtc Change quorum drive from q: to w: (right click on cluster, go to quorum tab, choose W: drive) Change msdtc log path to w:\ in MSDTC component services control panel Online msdtc, test failover. If failover is successful, disconnect legacy quorum Add new final quorum as Q: Repeat process to move to Q:\ Test cluster failover
Patrick,
In regards to the LUN serials changing, I've never seen this to be an issue with MSCS, but I've heard it can be a big gotcha when snapmirroring VMWare LUNs. Opinions?
- Hadrian
-----Original Message----- From: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Klise, Steve Sent: Friday, May 16, 2008 10:50 AM To: G.Milazzo@sinergy.it; pvh@databasement.eu; toasters@mathworks.com Subject: Re: R: MSCS and snapmirror question
Correct. Same filer. Thanks
----- Original Message ----- From: Milazzo Giacomo G.Milazzo@sinergy.it To: Patrick van Helden pvh@databasement.eu; Klise, Steve; toasters@mathworks.com toasters@mathworks.com Sent: Fri May 16 10:29:55 2008 Subject: R: MSCS and snapmirror question
Of course but that guy said he's migrating on the same filer to another aggregate.
________________________________
Da: Patrick van Helden [mailto:pvh@databasement.eu] Inviato: ven 16/05/2008 19.12 A: Milazzo Giacomo; Klise, Steve; toasters@mathworks.com Oggetto: RE: MSCS and snapmirror question
Guys,
Please be aware that MSCS will keep track of the lun serials. So you would probably need to set the lun serial numbers back to the original ones when your snapmirroring to another filer.
Regards,
Patrick van Helden Databasement BV pvh@databasement.eu
-----Oorspronkelijk bericht----- Van: owner-toasters@mathworks.com namens Milazzo Giacomo Verzonden: vr 5/16/2008 18:01 Aan: Klise, Steve; toasters@mathworks.com Onderwerp: R: MSCS and snapmirror question
You will not have issues. And you don't need to update your mirror. If you simply power off you clusters no i/o activity will be performer on LUNs. Then you can simply copy (vol copy) the volume(s) containing your LUNs, rename the new volumes as the old (in case of iSCSI LUNs and if you don't want to recreate the connections) and restart you clusters.
In case your nodes (but it will not happen) will loose the LUNs with Snapdrive you can safely reconnect them starting with the quorum one, on first node then restarting the cluster service, then on the second node and so on with all other cluster shared LUNs
Regards
-----Messaggio originale----- Da: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] Per conto di Klise, Steve Inviato: venerdì 16 maggio 2008 15.50 A: toasters@mathworks.com Oggetto: MSCS and snapmirror question
I am migrating from an aggregate to another aggregate in the same filer (6030). The LUNS are attached to print clusters, 2 node and 3 node clusters (WIndows 2003 sp2 x86 etc).
I have snapmirrored the volumes. If I stop the cluster service on the windows boxes, do a snapmirror update, the disconnected the quorum and other resource disks, do a snapmirror update, release the snapmirror, remove and replace the disks with snapdrive with the same disk letters. fire up the cluster service.. Will that work? has anyone done that before? Any gotchas?
Any help is greatly appreciate!
-
Hadrian Baron -
Klise, Steve -
Milazzo Giacomo