To add to dwilliams62, also be sure to get the latest host integration toolkit when using a 2008 R2 cluster (this may just be for the 5.0.2 firmware, but even if you initially stay with 4.3.7, you may later upgrade to 5.x, so it's best to already be on the newer hitkit.
Thanks for the document. I was looking for information like that, but couldn't find anything.
If I am reading it correctly, it appears this is the best practise setup:
1)VM on the local Hard Drive of the server (1 VM per drive if possible). The server in question has 4 * 146GB SAS...but it sounds like putting it in a RAID configuration is not the best idea. Each drive should should be "separate" (for lack of a better term).
2)A Data drive for each VM created on the SAN
3)When I am ready to cluster, I setup CVS, so the VM's move back and forth on the cluster (when needed).
That won't work, as a clustered resource needs to be completely located on shared storage (not the "local hard drive of the server").
Also, if you plan to cluster, but currently don't have a 2nd server, I'd suggest to make it a 1-node cluster with CSV and everything set up already. This way all you have to do later, is join the 2nd host to the cluster.
Microsoft's hyper-v isn't as capable with "importing" VMs are VMware is; in VMware ESX/ESXi you browse the datastore, right click the .vmx file and select "add to inventory". Microsoft doesn't have anything near this easy in hyper-v.
Dev Mgr
4 Operator
•
9.3K Posts
0
November 18th, 2010 14:00
To add to dwilliams62, also be sure to get the latest host integration toolkit when using a 2008 R2 cluster (this may just be for the 5.0.2 firmware, but even if you initially stay with 4.3.7, you may later upgrade to 5.x, so it's best to already be on the newer hitkit.
hutcha4113
1 Rookie
•
49 Posts
0
November 19th, 2010 06:00
Thanks for the document. I was looking for information like that, but couldn't find anything.
If I am reading it correctly, it appears this is the best practise setup:
1)VM on the local Hard Drive of the server (1 VM per drive if possible). The server in question has 4 * 146GB SAS...but it sounds like putting it in a RAID configuration is not the best idea. Each drive should should be "separate" (for lack of a better term).
2)A Data drive for each VM created on the SAN
3)When I am ready to cluster, I setup CVS, so the VM's move back and forth on the cluster (when needed).
Does that make sense?
Thanks again.
Dev Mgr
4 Operator
•
9.3K Posts
0
November 19th, 2010 12:00
That won't work, as a clustered resource needs to be completely located on shared storage (not the "local hard drive of the server").
Also, if you plan to cluster, but currently don't have a 2nd server, I'd suggest to make it a 1-node cluster with CSV and everything set up already. This way all you have to do later, is join the 2nd host to the cluster.
Microsoft's hyper-v isn't as capable with "importing" VMs are VMware is; in VMware ESX/ESXi you browse the datastore, right click the .vmx file and select "add to inventory". Microsoft doesn't have anything near this easy in hyper-v.