Long time to Rescan Disks

[compupartes]

Hi

I am having some issues with my new implementation of vSphere 5, this is my enviroment:

     1.- I have 3 Host with ESXi 5.0 update 1. BL465 G7 (ROM Firmware A19 03/19/2012, ILO3 Firmware 1.28 Jan 13 2012)

     2.- I have 1 P6300. (last firmware)

     3.- I have 2 Switches Brocade 4/12. (Firmware v5.3.0a)

     4.- Other server BL465 G1 and BL685 G1 (Windows 2003).

The problem is when I reboot one ESXi server it takes a long time to boot or when I do a disk rescan. I notice that the boot screen shows this messages for about 15 min each one:

     vmw_satp_alva loaded successully

     vmkibft oaded successfully

     vmf3 loaded successfully

   

and if I look at the vmkernel log I see these messages:

2012-09-27T19:39:27.004Z cpuO:8867)WARMIMG: Partition: 944 : Partition table read from device naa.6001438012599c460000600000400000 failed I/O error
2012-09-27T19:39:37.013z cpu1:8193) ScsiDeviceIO:2322: Cmd(0x4124001fe300) 0x9e, CmdSN 0x64 from world 0 to dev "6001438012599c4600006000004c0000" failed H:0x5 D:0x0 p:0x0 Possible sense data: 0x0 0x0 0x0.
2012-09-27T19:39:47.023z cpu1:8193) ScsiDeviceIO:2322: Cmd(0x4124001fe300) 0x1a, CmdSN 0x66 from world 0 to dev "6001438012599c4600006000004c0000" failed H:0x5 D:0x0 p:0x0 Possible sense data: 0x0 0x0 0x0.
2012-09-27T19:39:57.023z cpu1:8193) ScsiDeviceIO:2322: Cmd(0x4124001fe300) 0x28, CmdSN 0x67 from world 0 to dev "6001438012599c4600006000004c0000" failed H:0x3 D:0x0 p:0x0 Possible sense data: 0x0 0x0 0x0.
2012-09-27T19:39:57.023z cpu2:8867) Partition:484: Read of GPT header failed on "6001438012599c4600006000004c0000": I/O error
2012-09-27T19:40:07.023z cpu0:8194) ScsiDeviceIO:2322: Cmd(0x4124001fe300) 0x28, CmdSN 0x68 from world 0 to dev "6001438012599c4600006000004c0000" failed H:0x3 D:0x0 p:0x0 Possible sense data: 0x0 0x0 0x0.
2012-09-27T19:40:07.023Z cpuO:8867)WARMIMG: Partition: 944 : Partition table read from device naa.6001438012599c460000600000400000 failed I/O error
2012-09-27T19:40:07.029Z cpu2:8194)NMP: nmp_ThrottleLogForceDevice:2318: Cmd 0x1a (0x4124001fe300, 0) to dev "vmhba32:C0T0:L0" on path "vmhba32:C0:T0:L0" Failed: H0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x24 0x0. Act
2012-09-27T19:40:07.029Z cpu2:8194) ScsiDeviceIO:2322: Cmd(0x4124001fe300) 0x1a, CmdSN 0x6f from world 0 to dev "mpx.vmhba32:C0:T0:L0" failed H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x24 0x0

The message repeats for every RDM disk I have.

I think that this issue has to be with the SAN Switches, maybe I should update the firmware o maybe I have grong another configuration.

Could some one help me please?

Thanks.

[sparrowangelste]

whats the latency

maybe your san is not keeping up with the deamnds on it

": Partition table read from device naa.6001438012599c460000600000400000 failed I/O error"

[compupartes]

Hi thanks for reply.

I do not think so, It is not a heavy used enviroment.

As additional information, the disks that appear with problems are RDM that are used for a Physical to Virtual cluster.

I am looking at the "Setup For Failover Clustering and Microsoft Cluster Service" document and I just have some configuration Issues:

     1.- C: disk is in Thin Provision and I will change it to Thick provision eagerzeroedthick, no problem for that.

     2.- I am not using the driver STORport miniport for the physical machine, I will see if i can change it. 

     3.- Lastly, I do not understand it but I think that could be part of the pronlem, the document advise me to avoid the use of multipath software in physical and virtual machine. I am using it.

I hope to get help.

[Kahonu84]

Aloha,

We had a similar issue and once we did a clean ESX install, scan times reverted back to normal.

Bill

[compupartes]

It could be. Monday I will try with one server and let you know.

Thanks...

[mikemooney]

Having the exact same issue, did the ESXi OS re-install help or was it something else?

Thanks,

Mike

[compupartes]

It does not help, you should try this.

VMware KB: ESXi/ESX hosts with visibility to RDM LUNs being used by MSCS nodes with RDMs may take a long time to boo…

Is the KB 1016106

[grasshopper]

Hi Mike,

You are reviving a relatively old thread.  Keep in mind it's always best to create a new question and provide a link back to this one as a reference.  This way you can get better support and have the ability to award points to those that help you.

With that said, you might consider updating to the latest release (available as of last Thursday).  It addresses the following (among other things):

• PR1005316: In the presence of unresolved VMFS volumes (for example, VMFS volumes that are in unmounted state or VMFS volumes that are snapshots of other VMFS volumes), tasks that require storage to be rescanned or LUNs to be detached might take excessive time to complete. If ESXi hosts have access to many such volumes, these delays might cause tasks initiated from vCenter Server to time out, and consequently result in ESXi hosts getting disconnected from vCenter Server.

Best of luck and have fun!

-grasshopper

[mikemooney]

Thanks for the responses and apologies for my delay in responding, the KB posted by compupartes resolved the issue....

Very much appreciated...

Mike

[grasshopper]

Excellent!  Thanks for sharing.