Dear VMware Support,
I have just upgraded my ESXi 5.0 Update 1 to ESXi 5.1.
The host is an HP ProLiant DL360 G7 and the passthrough device is an HP PCI_E SC11Xe HBA (which actually is an LSI 53e1030 PCI-X Fusion MPT Dual Ultra320 SCSI).
I must say that the passthrough device is used to allow a VM to run backups using a SCSI tape and that the system was running fine with ESXi 5.0 Update 1.
After installing ESXi 5.1 I configured a fresh new VM with this PCI passthrough device and started it, but as soon as the VM boots (no OS installed yet) the ESXi host goes to purple screen (see the attached file for complete info):
VMware ESXi 5.1.0 [Releasebuild-799733 x86_64]
PF Exception 14 in world 25718:vmx IP 0x41802c0f095c addr 0x14
Code start: 0x4182be00000 VMK uptime 0:05:41:28.716
0x412211d9bcd8:[0x41802c0f095c]PCI_GetExtCapIdx@vmkernel#nover+0x2b stack: 0x412211d9bd38
0x412211d9bd48:[0x41802bebadd2]VMKPCIPassthru_GetPCIInfo@vmkernel#nover+0x335 stack: 0x230001700301
0x412211d9bef8:[0x41802c276791]User_LinuxSyscallHandler@<None>#<None>+0x17c stack: 0x41802be4cc70
This error occurs even if there's no device attached to the HBA and no OS on the VM. ESXi 5.1 crashes as soon as I turn on the VM (I suppose it is during the POST).
The server and the HBA have their firmware up to date.
This issue seems to be related to the latest release of ESX 5.1 and it is a regression from ESX 5.0.
Since I can easily reproduce the problem, please let me know you need any other info from my environment.
In the meantime I'll test the same configuration on another HP server (this time a ProLiant DL380 G7).
Thanks in advance
you probably need to open a support request with vmware for this since its a exception 14 hardware issue
actuall check this thread:
dariusd "We've isolated the problem and have an internal bug report open to track the fix.
The problem should (mostly?) only affect PCI devices as opposed to PCIe devices; "
looks like it relates to you
Hi Renato, and welcome to the VMware Communities,
Firstly, please note that the VMware Communities website is not an official channel to VMware Support. We have a page listing the official channels by which you may open a Support Request; refer to those instructions if you feel a Support Request is necessary.
We have isolated the cause of the issue you have encountered. It should only affect PCI (not PCIe) devices, although from your description you have a device with PCIe host interface and form factor which is layered around a PCI-X device, so it is understandable that it will also be affected.
We have an internal bug report filed to track the issue and the delivery of a fix. As sparrowangelstechnology has mentioned, you may consider opening a Support Request if you need more immediate assistance with your scenario.
I hope this helps!
Hi Darius and sparrowangelstechnology,
thank you for your very quick answers.
I have to admit that while diagnosing the purple screens, I tried to search for any reference on VMware Communities, but it seems I lost some posts, and when I wrote "Dear VMware Support" I actually knew I wasn't writing to the official product support and there are other ways to contact them.
Anyway, thank you for pointing that out.
Moving to the passthrough issue, I've currently removed the device from the VM so that there's no risk of crash for the host and I've reconfigured my backup software to use a different device instead of the tape attached to the SCSI HBA.
Now I'll start checking for upcoming fixes to ESXi 5.1.
Is it expected to read something about that on this thread or on the other one linked by sparrowangelstechnology?
Thanks again for your support.
The issue was fixed in the latest ESXi 5.1 patch that was released yesterday: