ESXI 5.1.0 purple screen with openwrt

[kutyeli]

Hello,

Very annoying problem. I run an openwrt instance and after a few reboots the ESXI 5.1.0 shows purple core dump screen on the console.

I have an HP Proliant DL 360 G7 with latest firmware update from HP. Also installed the latest HP drivers. It should be fully supported

but no cigar.  I did not encounter this problem with ESXI 5.0.0. Anyway around this problem? What kind of logging do you need from me to look

into this?

Regards K Yeli

[spravtek]

Hi, welcome to the community ...

What kind of error do you get in the purple screen? Can you post a screenshot of it?

To get more information about purple screens, if you haven't seen it already: http://kb.vmware.com/selfservice/search.do?cmd=displayKC&externalId=1004250

As for logs, the vmkernel.log or the vmkwarning.log would be interesting to start with ...

[kutyeli]

hello Spravtek,

Thanks for the reply. Please find attached the wanted data.

I hope it helps finding the problem. If there is anything more I can do please let me know.

I am about to downgrade 10 servers which is not a very nice job just before christmas! 🙂

Regards

K. Yeli

[spravtek]

Hi, quick question ... What is the exact patch-level/releasenumber of your ESXi servers?

[kutyeli]

Esxi 5.1.0 build 838463? Installed it fresh from the 5.1.0 DVD. Is this exact patch level somewhere available?

[spravtek]

No, that's what I needed, thanks ...

You installed ESXi 5.1.0a from the HP ISO or from VMware ISO and then installed the drivers separately?

Just to be clear on this ...

It could still be hardware/driver related ...

[kutyeli]

I tried both in fact. The version now installed is the HP ISO.

[spravtek]

Are you using pass through at all? ... Some reported problems with that as well ...

[spravtek]

There might be an issue with storage as well ...

From the logs

nmp_ThrottleLogForDevice:2319: Cmd 0x85 (0x4124403cdd40, 5260) to dev "naa.600508b1001c33be9a396e400afec6ba" on path "vmhba0:C0:T0:L1" Failed: H:0x0 D:0x2 P:0x0 Valid sense data: 0x5 0x20 0x0. Act:NONE

Some sort of flapping maybe ...

But it could also be that it is triggered by something else ... Maybe network cards ... What are you using as storage?

[kutyeli]

hi,

I use the USB passthru. I have an USB modem connected to the openwrt instance.

Therefore I have to disable the usb-storage. (esxcfg-module -u usb-storage).

But even before enabeling this passthru I get the resets.

Regards.

[kutyeli]

I use the native network cards on the main board of the HP DL 360 . There are 4 of them. No additional network cards. Do you want more details about this?

Storage is 2 SDD drives in HP hardware raid 1. Do you want model types as well?

If I can provide more answers please ask!

[spravtek · Accepted solution]

Well ... Normally this isn't for your build .. But the purple screen you have does look remarkably similar to this:

http://kb.vmware.com/selfservice/search.do?cmd=displayKC&docType=kc&externalId=2038767&sliceId=1&doc...

Does your ESXi come up/stay up at all or is it completely down? ... If you can access it, you could test with this setting ... Though as I said, normally if your build is higher than this it should be solved.

[kutyeli]

Dear spravtek,

I did the /BufferCache/MinZeroCopyBufferLength -i 512 change.

I have performed a number of restarts of the instance and the ESXI host remains stable now!

But when I close down the openwrt instance in the console by command halt and after that performing 'power down' in

the sphere client the host still reboots.

K Yeli.

[spravtek]

That's excellent news ... I hope the fix holds up and you can enjoy the holidays

Maybe that bug isn't fixed in the latest update, or not for all machines, thing to keep an eye on.

Thanks for the update!