ESXi 5.x "I/O Error Occurred"

[spoonuk]

Hardware:

Supermicro X8SIL-F-O with Xeon X3450 CPU

IBM ServeRAID M1015 flashed to LSI 9211 IT Mode

Onboard Dual Intel 82574L Gigabit Ethernet Controllers

HP NC360T 2 Port 1GbE NIC

Intel X520-DA2 2 Port 10GbE NIC

Samsung 830 256GB SSD

Seagate Barracuda 2TB [ST2000DM001]
Vertex 1 30GB SSD
Patriot Extreme Performance Xporter XT Rage 8GB (ESXi)

ESXi 5.0 but tried 5.0 U1, U2 and 5.1.

Please find attached error messages (2)

This issue is only happening when Seagate Barracuda ST2000DM001 is used as a datastore, I have two other datastores on Samsung and OCZ SSDs and there are no problems with them. Basically 10 seconds into file upload using vSphere client or migrating VMs between datastores I get "I/O Error Occurred" error message. Shortly after vSphere client cannot see the datastore anymore and it simply disappears! Attempts at re-creating it fail (see second screenshot) but the disk appears to be visible - I just cannot do anything with it. After a reboot the disk is completely gone and I have to shut down in order to see it again....

Windows Server 2008 R2 SP1 installed directly on the hard drive was not giving me any issues at all – it went through lots of testing and the drive is 100% good.

Did you guys come across anything similar?

I have seen plenty posts on the internet with users successfully using M1015 controllers and everyone seem to be very happy with them.

Help!

[spoonuk]

This cannot be right, surely?

~ # esxcfg-scsidevs -a

vmhba0  ata_piix          link-n/a  sata.vmhba0                             (0:0:31.2) Intel Corporation Ibex Peak 4 port SATA IDE Controller

vmhba1  ata_piix          link-n/a  sata.vmhba1                             (0:0:31.5) Intel Corporation Ibex Peak 2 port SATA IDE Controller

vmhba2  mpt2sas           link-n/a  sas.500605b002a638e0                    (0:2:0.0) LSI Logic / Symbios Logic LSI2008

vmhba32 usb-storage       link-n/a  usb.vmhba32                             () USB

vmhba33 ata_piix          link-n/a  sata.vmhba33                            (0:0:31.2) Intel Corporation Ibex Peak 4 port SATA IDE Controller

vmhba34 ata_piix          link-n/a  sata.vmhba34                            (0:0:31.5) Intel Corporation Ibex Peak 2 port SATA IDE Controller

~ #

~ #

~ # vmkload_mod -s mpt2sas |grep Version

Version: Version 06.00.00.00.6vmw, Build: 623860, Interface: 9.2 Built on: Feb 17 2012

Version 6 :smileyconfused:

I'm running:

~ # vmware -v

VMware ESXi 5.0.0 build-914586

Version 6 seems way too old...

[spoonuk]

I'm running the latest drivers available but the problem still persist.

~ # vmkload_mod -s mpt2sas |grep Version

Version: Version 15.00.01.00.1vmw, Build: 472560, Interface: 9.2 Built on: Nov 22 2012

Attaching VMKernel.log when the issue occurs (while copying 250MB .mkv file)

[mortengh]

Do you think this is the same or related to my issue?

http://communities.vmware.com/thread/445894

[loiphin]

Yip, just spent the best part of 7 hours f00king about with my ESXI setup and wondering why it took twice as long to do a installation of Win2012 server on a ST2000DM001, as compared to an older Samsung F1 drive. I tried every imaginable combo and eventually realised the drive is duff. Hence me landing up at this post

I ended up flashing the latest firmware to the ST2000DM001, but it made no difference. And I tried the same model HD out of my gaming rig and both end up working like sh!te on ESXI 5.1 U1.

Just popped in an older Seagate Barracuda 1.5TB drive and it also works just fine. So something is screwed with the ST2000DM001 on ESXI. On my Win8 rig its just fine!?!?!?

loiphin.