Purple screen on ESXi 6.5.0, 6765664

[btesrail]

Dear experts,

Recently, I got several purple screen on our running production ESXi 6.5 server.

Version: DellEMC-ESXi-6.5U1 VMware ESXi 6.5.0, 6765664

Server: Poweredge R740

CPU: 2 x Intel Xeon Silver

Memory: 64GB

Anybody faced the same and any resolution?

Thanks.

The purple screen was attached.

[MBreidenbach0]

Google reveals this: VMware Knowledge Base

There are similar search hits with other NICs,

So it MAY be NIC firmware / driver.

Just a guess.

[btesrail]

Hi Guru,

Thanks.

The server has integrated 4-ports 1GB NIC.

I added another dual-port 1GB NIC by PCI-E.

The server has been running for 3 months and no issue.

Recently, it got the purple screen several times.

I am trying to adjust virtual networking and removing this NIC to eliminate possible reason.

BR,

Eric

[btesrail · Accepted solution]

The problem has been resolved and after monitoring several days, no purple screen anymore.

Below is my resolution:

1.Check performance chart/log of esxi host and VMs. Noticed that the upper utilization achieve 100% before purple screen happened on several cpu cores.

2.Identify CPU utilization of VMs and found out that one Windows server 2008 R2 std VM server (2 vCPUs) had 100% CPU utilization on one core for some time. Then the performance chart lost. Checked the time when chart lost and it was the time esxi host got purple screen.

3.Review CPU allocation to all VM servers, especially the VM identified above. Adjust number of vCPUs  of above VM server to 4 vCPUs.

Until now, no purple screen and CPU abnormal performance anymore.