Last December (2016), I was deploying a 2-node VMware ESXi 6.5 cluster on 2 brand-new HPE ProLiant DL380 Gen9 rack-mount servers for a customer as a P2V and HA project.
A custom HPE image was used for deploying ESXi 6.5.
The cluster was working fine for a few days but later started throwing Purple Screen Of Death (PSOD) or Purple Diagnostic Screen stating unable to acquire lock with a few backtrace details on a member server.
Tried searching for solutions on the VMware Knowledge Base articles but found none.
In late January (2017), HPE released a Customer Advisory related to this issue.
The fix was to update the iLO driver to the latest version or at the least higher than 6220.127.116.11.
Recently (August, 2017), VMware also came out with a KB for the same.
You can find your iLO driver version by logging onto the HPE iLO 4 web console and go to Information > System Information > Software > hpe-ilo.
You can download the iLO Driver vib from HPE vibsdepot (latest available as of Sep 24th, 2017, day of writing this post, is: ilo_618.104.22.168-12.4240417.vib), place it on a datastore accessible to the affected ESXi Host and then install it.
HPE Customer Advisory c05378386: http://h20564.www2.hpe.com/hpsc/doc/public/display?docId=c05378386