Need suggestions for SAN migration

[dustin.wooldridge]

A client we currently manage has decided to upgrade their existing 2 SAN's (OLD01/02) to 2 new SAN's (SAN01/02) with faster drives. OLD01-02 are running Starwind Version 5.6 and are configured with HA images connected to HYPER-V01/02 via iSCSI (native Windows iSCSI initiator). OLD01 replicates data to OLD02. OLD01/02 also provides DFS storage as well as backup storage for VM's using Backup Exec 2010R3. The performance has been very very poor (not Starwind's fault, the SAN's just do to much and have slow disks), to the point of unusable at times. The new SAN's will have 36 600GB 15K RPM SAS drives, and I would like to use RAID 10 instead of RAID5 which is currently being used on the old SAN's. Also, the OLD SAN's I believe only have 1 large RAID5 array configured, we will not be doing this on the new servers.
We would like to install Starwind 5.8 onto the new SAN's, but I would like a little bit of help on the best way to migrate the HA/CSV images to the new SAN's, as well as any pointers you might have.
Thanks a bunch!

The current setup is attached as well as the setup we would like to use once the upgrade is completed.

new setup.jpg (39.71 KiB) Viewed 7866 times

old setup.jpg (39.54 KiB) Viewed 7866 times

HYPER-V01/02 details:
Dell PowerEdge T710 dual Xeon X5550 72GB RAM
Server 2008 R2 SP1
Using NIC teaming

OLD01/02 details:
SuperMicro server running Storage Server 2003
Using NIC teaming

SAN01/02 details: Don't have all the details on the hardware yet, server is due to arrive in a week or so
Server 2008 R2 SP1

[Anatoly (staff)]

Deart Dustin,

That are the good plan, good HDDs and good RAId choice!
I think the best way will be as folows:
1. To prevent data loss, disconnect clients from HA;
2. Install 5.8 on the both SAN machines (old ones);
3. Copy *.img and *.swdsk files from old SANs to corresponding new machines;
4. Install 5.8 on new machines;
5. Create HA devices using copied img files (you can choose "do not synchronize" option they to appear synchronized);

After HA devices will be created you can connect to them and use the data on them.

[dustin.wooldridge]

Anatoly,
Sorry for the late reply, but I just wanted to thank you for your suggestion. We'll be performing the upgrade this month and I'll be using the method you supplied.
Thanks!

[Anatoly (staff)]

No problem!
I hope everything will run smoothly!
We will look forward to hear back from you!

[dustin.wooldridge]

Anatoly,
We ran into some issues....
When the old SAN's were first setup, the network adapters for iSCSI traffic and sync were not setup correctly. Actually, no dedicated sync channel was created in the first place. Aside from that, our HA images (4 HA images 1 HA-Quorum, 4HA-Partner images 1 HAQuorum-Partner) are 1TB in size each and the file transfer (using teracopy) was estimated to take 7 Hours to complete. That's just too long for our network to be down.
Do you have any other suggestions for getting the old Starwind images onto our new SAN's without 28+ hours of downtime?

Also, I had to remove the 4 port iSCSI team we initially used on the old SAN's so that I could use 2 of the ports dedicated to sync and the other 2 for iSCSI traffic. After I removed the team, Starwind would not let me change the sync and heartbeat interfaces on the original images.

[Anatoly (staff)]

Well, there should be one.

You need to upgrade you 5.6 (SANold1 and SANold2) first and install 5.8 (SANnew1 and SANnew2) on new boxes. After it`ll be done you can use Change Partner feature, for example on SANold1box (right click in the device and choose corresponding option), therefore SANold1 and SANnew1 will start functioning as HA storage. After the synchronization will be finished right click on device SANnew1 box (for this example) and choose change partner option, and choose SAN2 box to was a new partner for SANnew1 server.

[dustin.wooldridge]

Anatoly,
Thank you for your suggestion. It's funny how we can overlook the most obvious of things when in the heat of the moment . We were able to work everything out and get the client back up and running.
Luckily, I shutdown all the VM's and disconnected all sessions before making any changes- which is good since I accidently initiated a full sync which could have proven disasterous had any changes been made to a the partner node. Thank goodness for scheduled maintenance. I have a support call later this afternoon to check up on some read/write performance issues I was having when connecting to both HA and HA-Partner. Other than that, I'm getting great speed's on the new SAN's and really loving the ability to increase the size of images on the fly. This is going to save us a ton of space and allow easier scalibility in the future.

Thanks!

[Anatoly (staff)]

Human factor - 8th Wonder of our World

OK, I`m glad that our assistance appeared to be helpful for you!

We`re always opened in case of need.