70 minute timeout deployng template to local storage

[Jae_Ellers]

Anyone seen a 70 minute timeout anywhere in VC or on the ESX host?

We're trying to deploy a std 2003 template to local disk on a test server. The server is an old DL380 G3, probably without BBWC and so takes a long time to copy the 15G template in.

After 70 minutes (89%) and some change I get an Operation timed out error in VC and the deployment stops. However, the disk continues to deploy and the vmdk file ends up in the right place, but with no config.

I'm deploying with the guest customization wizard.

VC 2.0.1p2, ESX 3.0.1 current patches.

The only timeouts I see in VC are Client Connections, which are set to 60 seconds and 240 minutes.

I've tried boosting the following based on some forum searches, but they didn't help:

Scsi.SCSITimeout_ReabortTime from 5000 to 50000

Scsi.SCSITimeout_ScanTime from 1000 to 10000

vpxd.log

Task starts

\[2007-09-19 07:40:30.947 'App' 4552 info] \[VpxLRO] -- BEGIN task-12897 -- vm-20434 -- vim.VirtualMachine.clone

I see a few of these

\[2007-09-19 08:49:28.589 'Locale' 5204 warning] Message file access using key 'vm.guestos.winNetStandardGuest.label' with no default supplied.

Then the task fails:

\[2007-09-19 08:50:07.293 'App' 4552 error] \[VpxdVmprovUtil] Unexpected exception received during NfcCopy

\[2007-09-19 08:50:07.527 'App' 4552 error] \[clone] (tek2003sp2 20070918) Unexpected exception (vim.fault.Timedout) during clone. Aborting.

\[2007-09-19 08:50:07.527 'App' 4552 info] \[VpxdVmprovXaction] Starting journal rollback

\[2007-09-19 08:50:07.527 'App' 4552 info] \[VpxdVmprovXaction] Undo GenericTask... Host: host-6144 ,Task: task-15159

\[2007-09-19 08:50:07.683 'App' 4552 warning] \[VpxdVmprovXaction] Timedout canceling task on host

\[2007-09-19 08:50:07.855 'App' 4552 info] \[VpxdVmprovXAction] Undo CopyVmFiles... Host: host-6144 ,filesCopied: sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/v-jamesel-test1.nvram, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/vmware-7.log, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/vmware-8.log, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/vmware-6.log, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/vmware-9.log, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/vmware-10.log, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/vmware-5.log, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/vmware.log, sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/v-jamesel-test1.vmdk,

\[2007-09-19 08:50:08.308 'App' 4552 info] \[VpxdVmprovXAction] Undo CreateInvtVm... vmPath: /vpx/vm/#20437/ ,inDb: true ,onHost: false ,onDisk: false

\[2007-09-19 08:50:08.386 'App' 4552 info] \[VpxdVmprovXAction] Undo ReserveName... Host: host-20417 ,Name: sanfs://vmfs_uuid:46c101e5-50092662-caf9-000bcd294e1f/v-jamesel-test1/v-jamesel-test1.vmx

\[2007-09-19 08:50:10.215 'App' 4552 info] \[VpxdVmprovXAction] Undo CreateInvtVm... vmPath: /vpx/vm/#20437/ ,inDb: false ,onHost: false ,onDisk: false

\[2007-09-19 08:50:10.215 'App' 4552 info] \[VpxdVmprovXaction] Vm provision transaction rollback done

\[2007-09-19 08:50:10.215 'App' 4552 error] \[VpxdVmprov] Clone failed: vim.fault.Timedout

Any ideas? We'll open an SR, try this on a faster system, etc, but we'd like to be able to use local disk for some testing.

[boydd]

Have you tried restarting the management agent on the host and then trying?

DB

[Jae_Ellers]

Yep.

[Jae_Ellers]

Turns our this is a known issue when copying from fast storage (SAN) to slow storage (local disks).

Turns out we didn't have bbwc on the test systems, so we ordered a couple and saw a huge increase in write performance (of course).