#PF error diagnosis - Can anybody help?

[Dent13]

I've started getting regular PSODs.  I'm trying to figure out the problem.  They happen after about 7-12 hrs uptime.

I am using VMware ESXi, 5.5.0, 2068190  from (Updated) HP-ESXi-5.5.0-Update2-iso-5.76.36 with EVC set to Merom.

Both stack traces start with E1000TxTSOSend - Am I looking for NIC issues here?  Not sure where to go from here...

[jonathanp]

Hi,

    if you do have VMware support, create a support bundle and open a case.... that would be the fastest way....

Otherwise...you can look at few things...

1st.... you error is a Page Fault... ( 1st line under the Product Version....)

#PF Exception 14..

See KB

VMware KB: Understanding Exception 13 and Exception 14 purple diagnostic screen events in ESX 3.x/4....

Also you cann look at the logs on ESXi in "/var/log/*"

look at around the time of failure, you may find some hint...

There is also a KB on how to interpret the data on a Purple Screen....

VMware KB: Interpreting an ESX/ESXi host purple diagnostic screen

Hope that this helps...

jonathan

[FritzBrause]

Upgrade to latest NIC drivers.

And check http://kb.vmware.com/kb/2059053.

Of course you have U2 already installed.

But maybe you can switch VMs from E1000 to vmxnet and see what happens.

[Dent13]

jonathanp:

That's exactly what I did before I came here   Unfortunately, no support

[jonathanp]

I do not see the release due to the iLO banner at the top of your screenshot...

Are you at 5.5 Update 2....?

Jon

[Dent13]

It's in the text - 2068190

I am downloading this for a possible fix: VMware KB: VMware ESXi 5.5, Patch Release ESXi550-201410001

My fingers are crossed.

[Dent13]

Upgraded to latest patch release, but same problem.  Interestingly, it happened much more quickly on the new updated (one or two hours vs around one day).

I've narrowed it down to a single VM (OSX 10.10).  I switched from E1000 nics to E1000E nics and I've been good overnight.