Convert a physical machine but split data to 2 data stores

[ewebster]

I will be migrating a terminal server to vmware and it has three drives a C:, 😧 and E: drive.I was wondering if anyone knows how to place the C: and 😧 drive on datastore 1 and the E: drive on data store 2. Datastore 1 is located on the local machine that esxi is installed on and datastore 2 is a NFS share located on a NAS. If any one has done this or knows how to using vmware converter or any other way let me know thanks.

[kac2]

you can do this with storage vMotion.. but why would you ever use local storage? you won't have HA or vMotion abilities

[ewebster]

I am doing this conversion for a small multinational corperation and they enjoy low budget IT. so what we have si a vmware certified 8tb raid6 Qnap809Upro. And we were not sure even with the dual gigabit nics if we would have performance problems with puting everything on it.

[MattG · Accepted solution]

Using Standalone Converter, you can choose to put each volume/drive on a different datastore during the conversion.

I think you need to choose Advanced on the Datastore portion of the wizard.

-MattG

If you find this information useful, please award points for "correct" or "helpful".

[kac2]

you're running production ESX servers with only 2 NICs??? yikes... sounds like they don't know how a production internal cloud should function.

i would prefer storage vMotion...

other possible way. (I would take a snapshot or clone the VM before giving this a shot )

Shutdown the VM. Edit settings of the VM and Remove (do not delete) the virtual hardrives. COPY the .vmdk and vmdk_flat files of the drives to their new destination. once copy has completed. Edit settings of the VM, click add new hard drive from file, browse the datastore and find the folder of where you copied the VM, and add the VMDK to the VM. power on the VM and it should work.

i think Matt G probably has the easiest answer.

[golddiggie]

Do they realize that the TS-809 Pro and TS-809U-RP are both NOT VMware certified products? The TS-859U-RP IS "VMware ready" certified...

Depending on how the LUNs are created/carved up for the storage, and how you set the two NICs to behave, you could get decent performance out of the array. Don't expect it to be anything like what you'd get from a Dell MD3000i or EqualLogic array though.

Going below bargain basement for storage in any sized multi-national corporation is just moronic. I know of small, US based companies that spent the money to get proper storage, for their VMware environments. Doing such things right the first time actually saves you a lot more down the road. Such as when you need to add more arrays just to get mediocre performance since you cannot put the load on them that you would with proper storage solutions. Not to mention the additional man hours that will be spent administrating the low end solutions, which could be better spent on other things. Hell, for <$40k (if I recall correctly) you can get an EqualLogic PS6000 array populated with 1TB SATA drives. With the dual quad port ethernet controllers there, you wouldn't have an issue. Not to mention the stronger drive controllers, and more.

If they're not interested in spending the right money for storage for VMware hosts, I have to wonder how (severely) underpaid YOU are...

VMware VCP4

Consider awarding points for "helpful" and/or "correct" answers.

[ewebster]

This is just for a small branch of the corperation (5-10 users). they are however near the international terminal at the airport i think you can see it from their office. and I agree that the storage solution they chose is far below what they should be implamenting but they would rather have cheap and mediocre then expensive and sustainable.

[golddiggie]

How many host servers? How many physcial servers are being converted? What are the ESX servers? If just one server, depending on what it is, you would have probably been better off just installing a compliant SAS RAID card and using 1TB SAS drives inside the host... You would probably see better response times to the datastores that way, since it's not hitting the bottlenect at the crap array.

Or just put it all on the un-certified array and tell them that because they went so cheap, that's why performance sucks. IF they had bothered to spend a few dollars now, they would have had much better performance (possibly better than the old physical servers had). Plus, it's an item you buy once and not need to replace it later due to such poor performance. IF they need more space later, stack another on top of the existing, adding more space, boosting performance at the same time.

How much drive space do you actually NEED at that location?

Things like this is what happens when people in finance dictate what IT can purchase for solutions. Then it's up to IT to make the bottom run crap work properly, or to the same level as the quality products... If they factored in all the man-hours you, and your colleagues, will be spending on the products, the cost savings pretty much evaporate... Plus, if there's an issue with one of the drives inside the array you're being forced to work with, you have to contact at least two different manufacturers to figure out the issue, with unknown lag time between discovery and issue resolution (X days)...

VMware VCP4

Consider awarding points for "helpful" and/or "correct" answers.