See attached.
Fresh install of ESXi 4.1.0, trying to add a local datastore. Local volume is a 4.55TB RAID 5, fully initialized.
Add storage
choose the device, etc. Choose 8MB block size, and "mazimize capacity". Summary screen looks like its going to do what I ask...I click finish.
End product is 558.75GB. Any ideas?
ESX(i) does not support LUNs larger than slightly less than 2TB. Break your array into smaller LUNs
ESX(i) does not support LUNs larger than slightly less than 2TB. Break your array into smaller LUNs
If your controller supports virtual disks on an array you may be able to break the current array in to smaller chunks again slightly smaller than 2TB (2TB-512bytes). This is something at the controller level not ESXi virtual disks.
can i combine them again at the ESX level using the extend feature or do I need to make a span inside my guest OS?
Fresh install of ESXi 4.1.0, trying to add a local datastore. Local volume is a 4.55TB RAID 5, fully initialized.
I am sure your next question is "Yes, but why does it allow you to create 558.75GB partition"?
The answer is its a weird quirk of the limitation, it won't let you use the entire size, but it will let you use the REMAINDER, hence 4.55 TB - 4TB = 558.75
so if you break it up, there is a trick. Create a 1.9 TB data store in the Hardware RAID configuration (stay with me now, this get's good). Create another volume of equal size, then use the remaining space as a 3rd volume.
Now go into VI client, allow to find and create a datastore on that first volume, leave the rest for now. Once its up, name it to whatever you want. NOW the cool part, is ADD an EXTENT.
Keep adding extents to it until all the space is used up.. when you are done you should have around a 4.48 TB datastore.. voila, ALL the space in one datastore...
so how do shops with huge multi terabyte datasets get around this I wonder?
You can combine the Extents to create larger datastores. ESX(i) supports 32 Extents so you are able to create a 64TB datastore (minus all . . .)
When you have very very large datasets it is probably not a good candidate for virtualization.
Is this a question or just a post suggesting you have achieved success.
a post suggesting you have achieved success in helping me achieve success.
Yes, it worked. Broke it up into three equal chunks of 1551GB per virtual disk, created three datastores, then married em all up.
worked exactly as you said.
Success is always a good thing. Have fun.