Hi
I have one Datastore , Drive type SSD , after installing a windows 7 guest the OS doesn't detect the drive type as a SSD . Is there something im missing when im creating the VM Guest ??
Server is Esxi 5.5
Virtual machine type 8
Vmfs 5.60
Datastore is a local non shared raid 10 SSD array
Thanks Jamie
The disk will always Virtual Disk independent of what storage is presented at the back-end.
The guest OS should detect the virtual disk as a SSD , "ESXi allows operating systems to auto-detect VMDKs residing on SSD datastores as SSD devices" The OS will then be customised accordingly.
This is not the case with my win 7 guest
have you checked inside the OS? The screenshot you have pasted in on VM level.
Go to device manager and see if it has detected as SSD. Thats what documentation says. you are looking at the VM level.
Yes , I have checked the Guest OS , the virtual disk is not detected as a SSD as per my initial question (on windows 7 the disk defragmenter would be disabled for the SSD device)
Can you take a SSH session to your ESXi and run this command
cat /vmfs/volume/datatstore/<Win-7-vm>/<Win-7-vm>.vmx | grep scsi
You should see an entry as scsi0:1.virtualSSD=1 (Where scsi 0:1 is oyur SSD disk SCSI ID)
Let me know what it says.
scsi0.virtualDev = "lsisas1068"
scsi0.present = "TRUE"
scsi0:0.deviceType = "scsi-hardDisk"
scsi0:0.fileName = "Win7.vmdk"
scsi0:0.present = "TRUE"
scsi0:0.redo = ""
scsi0.pciSlotNumber = "160"
scsi0.sasWWID = "50 05 05 6e 11 09 c7 b0"
Is that the whole output?
Is there no line scsi0:1.virtualSSD?
Yes , that's the whole output
Btw how did you create this virtual SSD?
Hope this link helps you in this regard
-SatyS
Maybe you could try to upgrade your VM-version to 9 or 10 (backup first, of course), even if that article says it should work with VM-version 8. Or do you have any reason for still using version 8 (I see on the screenshot)?
BTW, I think it is actually useless to present virtual drive to VM as "SSD" if you are using local raid10-ssd datastore. No SSD-related command (i.e. trim) is passed to real SSD through raid-controller. AFAIK, this can be done only in pure raid0-configuration (and no vmfs in between), and even then not all raid-controllers and SSDs support it. VM-OS could maybe at least aling partitions to SSD-boundaries correctly, but then again you have vmfs in between which should take care of it first.
So as it is written in that virtuallyGhetto-article: you could emulate SSD in VM to see what it makes. But this is maybe useful for developers of SSD-related software. From practical point of view, this does not change anything with respect to underlying SSDs. Any other SSD-related optimization (i.e. turn indexing off, use ramdisk for temp to minimize writing, etc). you can do even without exposing storage as SSD to your VM-OS...