ESX 6.5 on HP DL380g7 – Fix PSOD

A word of precaution – the HP DL380g7 is not on the VMware HCL (Hardware Compatibility List) for ESX 6.5. It is ill advised to use ESX on unsupported hardware unless it is not for production use.

After installing ESX 6.5 on the host the kernel would panic and show the PSOD (Purple Screen of Death) with an error like: Exception 14 in world #####:sfcb-smx. Many people are reporting this same issue.

The fix I am documenting here is quick and dirty Рa more proper but time consuming fix is here.

Since this is unsupported and just for testing, I took a different path. Our issue here is with the hpe_smx kernel driver. When the init script runs at the final boot stage, it interfaces with hpe_smx and causes a kernel panic.

As an easy fix, we are going to remove hpe_smx from the kernel module configuration. This configuration is done inside boot.cfg, the boot loader configuration file.

To edit this file, I booted a Gentoo Linux minimal livecd, used ‘fdisk -l’ to list the partitions. I then mounted each one at a time to find which contained the boot.cfg file (mkdir /mnt/vmware, mount /dev/sda4 /mnt/vmware, ls /mnt/vmware, etc.) When I found it, I simply removed the entry for hpe_smx, and rebooted.

Once this is done, the issue should be resolved. I would recommend making a backup of the boot.cfg file before making changes (cp /mnt/vmware/boot.cfg /mnt/vmware/boot.cfg-orig).

Advertisements