I am having inconsistent performance with VMware 5.5.2. With a Windows 2003 Std Svr as host and Windows 2003 Svr as guest, I'm trying to use the SCSI tape system on the host box. I can see the Tape unit; Both NT backup and ArcServe on the guest can see the tape. However, when I insert a tape and start a job in NT backup, I get:
VMware Workstation unrecoverable error: (vmx)
NOT_IMPLEMENTED C:/ob/bora-29772/bora/devices/scsi/scsiPassthruWin32.c:683
which kills the guest machine.
There has been success with stripping 2K3 of all service packs. Does anyone have a clue which one is causing the problem or what patch would fix this issue?
I have a similar issue with VMware Workstation 6.5.1 build-126130 and HP SureStore DAT24x6 on Windows 2003 Enterprise Server.
The firmware of DAT24x6 device is latest revision U812.
The OS recognizes correctly the Changer C1557A and the Drive C1557A/HP.
However, when I started a backup job with Backup Exec, I got the following fatal error, then the VM dies:
> VMware Workstation unrecoverable error: (vmx)
> NOT_IMPLEMENTED d:/build/ob/bora-126130/bora/devices/scsi/scsiPassthruWin32.c:593
Looking into the vmware-vmx-5668.dmp, I didn't get much because it's just mimicking a system fault to gather system information.
Looking into the vmware.log, I could find the following log entries:
> Jan 06 00:54:54.078: vmx| SCSI PASS-THRU (scsi0:2): requested 4096 returned 32
> Jan 06 00:54:54.078: vmx| SCSI PASSTHRU: Successful operation w/ non-zero residual 4064
> Jan 06 00:55:22.078: vmx| scsi0:2: Command MOVE MEDIUM took 27.997 seconds (ok)
> Jan 06 00:55:22.093: vmx| SCSI PASS-THRU (scsi0:2): requested 4096 returned 32
> Jan 06 00:55:22.093: vmx| SCSI PASSTHRU: Successful operation w/ non-zero residual 4064
> Jan 06 00:55:22.125: vmx| SCSI PASS-THRU (scsi0:2): requested 1024 returned 324
> Jan 06 00:55:22.125: vmx| SCSI PASSTHRU: Successful operation w/ non-zero residual 700
> Jan 06 00:55:28.187: vmx| SCSI PASS-THRU (scsi0:3): requested 1024 returned 43
> Jan 06 00:55:28.187: vmx| SCSI PASSTHRU: Successful operation w/ non-zero residual 981
> Jan 06 00:55:28.234: vmx| SCSI PASS-THRU (scsi0:3): requested 1024 returned 43
> Jan 06 00:55:28.234: vmx| SCSI PASSTHRU: Successful operation w/ non-zero residual 981
> Jan 06 00:55:31.218: vmx| scsi0:3: Command REWIND took 2.976 seconds (ok)
> Jan 06 00:55:31.218: vmx| scsi0:2: Command TEST UNIT READY took 2.282 seconds (ok)
> Jan 06 00:55:35.765: vmx| SCSI PASS-THRU (scsi0:3): requested 65536 returned 8192
> Jan 06 00:55:35.765: vmx| scsi0:3: Command READ(6) took 4.548 seconds (ok)
> Jan 06 00:55:35.765: vmx| scsi0:2: Command TEST UNIT READY took 3.829 seconds (ok)
> Jan 06 00:55:35.781: vmx| NOT_IMPLEMENTED d:/build/ob/bora-126130/bora/devices/scsi/scsiPassthruWin32.c:593
> Jan 06 00:55:35.781: vmx| Win32 object usage: GDI 25, USER 24
> Jan 06 00:55:35.781: vmx| CoreDump_CoreDump: faking exception to get context
It seems that SCSI PASSTHRU driver of VMware has not-yet-implemented SCSI emulation.
Does anyone have an any information to solve/evade this issue?
I'm experiencing similar issues with VMWare server 1.0.9 and a Sony LIB-81 SCSI changer driven with Backup Exec 11d. Host OS is Windows Server 2003 Datacentre SP2, Guest is Windows Server 2003 Standard SP2.
"Apr 08 19:35:28: vmx| NOT_IMPLEMENTED C:/build/ob/bora-156507/pompeii2005/bora/devices/scsi/scsiPassthruWin32.c:683"
At which point the VM freezes and the console goes non-responsive. All I'm doing in Backup Exec is an inventory of the changer.
This is clearly an issue in many VMWare products. I don't much like the solution of rolling back to no service pack as alluded to here: http://communities.vmware.com/thread/94821
Is there any chance this can be looked into?