Link to home
Start Free TrialLog in
Avatar of aleghart
aleghartFlag for United States of America

asked on

ESXI & SSD - drive disappears

first time evaluating ESXI 5.  Installed on a Dell R210 on 4GB USB flash drive for ESXI.  Installed Crucial m4 SSD for storage of VM.  Worked OK during 2 days of testing.  _Very_ fast OS installation and boot up.

After OS installed & patched, make a snapshot called "before apps".
Then, started to install apps for users.
After a few items installed & tested, wanted to delete the snapshot.  Took 5 minutes to get to 25%.  10 minutes to get to 50%.  Stayed at 75% for over an hour, then I had to leave.  I left it to run overnight.

This was a <30 of 40GB used on the VM guest's drive.  It was the only VM on a 256GB SSD.

A 500MB application installation took about 1-1.5 minutes.  Why did this deletion of the snapshot take so long.

This morning, the VM is "unknown (inaccessible)".  I was going to assume it was toast, and figure out how to wipe/reprovision the SSD & start over again...but the SSD is no longer listed as a datastore.  The only things availble to me are a SATA HDD and an NFS store where I put ISO images for the VM guests.

What did I do to kill the SSD or take it offline?  Is it likely that the ESXI installation is trashed, or that I trashed the SSD?  Or a hardware failure not caused by me.

I'm not heartbroken over losing either the ESXI install or the VM guest...it was a first run at testing.  But, I'm lost as how to troubleshoot.
Avatar of Andrew Hancock (VMware vExpert PRO / EE Fellow/British Beekeeper)
Andrew Hancock (VMware vExpert PRO / EE Fellow/British Beekeeper)
Flag of United Kingdom of Great Britain and Northern Ireland image

Your SSD has had a possible hardware failure.

Check with Crucial if it requires a flash update to meet the specifications of your storage controller.

We've had approximately 24 SSD fail at present, including OCZ Revo Drive PCI-E cards.
Personally, I would not install ESXi onto an SSD to costly storage, especially when ESXi just boots the code into memory.

I would install ESXi onto a USB flash drive or SD card.

Here is the VMware KB on installing 4.1 on USB/SD:
http://kb.vmware.com/kb/1020655

Here is the VMware KB on installing 5.0 on USB/SD:
http://kb.vmware.com/kb/2004784

I would then just use the SSD for Datastore or SWAP.
If you had installed the ESXi installation on USB as stated, your ESXI installation is safe, but your SSD datastore, appears to have had a failure.

Is it still visible on the SATA controller at BOOT-UP?

Can you run any Crucial utilities to check if the SSD is okay?

Do you know how large the snapshot was? These can take many hours to remove depending upon size and speed.
Avatar of aleghart

ASKER

Am I better off at this point using Server 2008 and OS RAID 1?

No fancy controller...just the onboard SATA.
RAID 1 two disks SATA, performance IOPS, will be very poor.

Each SATA 7,200 rpm disk, is going to only give 50-80 per disk!
ASKER CERTIFIED SOLUTION
Avatar of Andrew Hancock (VMware vExpert PRO / EE Fellow/British Beekeeper)
Andrew Hancock (VMware vExpert PRO / EE Fellow/British Beekeeper)
Flag of United Kingdom of Great Britain and Northern Ireland image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
I was thinking RAID 1 with 2x 256GB SSD, not using 7200rpm drives.

Although, for the light load of terminal server, we're doing fine on an old 1.5Gbps SATA drive, installed natively on a PentiumD server.

I figured with wear on the SSD, I could still get 2 years on a heavily-hit drive.  5 years on a moderately used drive like a terminal server and print server, etc.

Next step:  full power off.  Now, the SSD shows up again.  It didn't after reboots.  Not sure what to think from this turn of events.

I created a new VM. Pointed it to the existing drive.  Booting to "Windows Error Recovery" asking if I want to boot normally.

I'll have to see where it sits with the snapshot.  Do I have apps, or reverted back to snapshot point "before apps".

This doesn't make me feel any better about running as a VM.  It's more fragile now than installing straight onto bare metal server.
OK, didn't boot normally.  Got the Windows splash screen, but shortly got a black screen with no progress.  It should have booted up in a about 30 seconds.

Trying "safe mode".

Nope.  Goes to black screen.

I believe the VM guest is toast.
Weve been seeing more issues lately with SSD, even on physical servers, e.g Win2k8 installation, quick install to SSD, completed all Windows updates, shutdown and restarted, it sat configuring updates for 24 hours, restart the server and the OS was toast.

We still do not understand why, using Corsair Force SSDs and OCZ!

We have now moved all our SSDs into our storage servers and present iSCSI and NFS to ESXi.
Using ZFS with the SSDs?
Yes, ZFS with SSDs, and ZFS L2ARC cache SSDs, ZIL - SSDs

We've also got standard 7,200rpm drives ZFS, but using L2ARC and ZIL - SSDs, and Logs.

Fantastic speeds, for NFS, iSCSI shareds storage.

Faster IOPS via iSCSI than using local SATA disks in servers!

We've been taking standard HP ProLiant MicroServers (£84 in the UK after cashback) and turning them into SuperSSD SANs for use with ESXi! (and they run at 64 Watts of electric!)

Ok...getting off-topic, buy how did you get the 6 SSDs attached if the 4 onboard SATA ports ate still available?

I may have to abandon local SSD if the are so troublesome.
LSI SAS 9211-8i

or

LSI SAS3081E-R

(but the MicroServer can have up to a total of 6 SATA in use, from it's motherboard, 4 on the Mini-SAS, and additional 2 - SATA on motherboard, 1 - esata on motherboard)

I think I'll test install OpenIndiana or Solaris Express 11.  Not sure how they will handle monitoring the RAID (alerting, etc.).  Can't flash the drive lights like I'm used to with older servers & SCSI RAIDs.

I may post another Q about straight OS  versus  ESXI with controller/drives passed through to a VM.