There is a bug that only affects ESX classic hosts running a specific
version of the Cisco Nexus 1000V software without using software iscsi
adapters. For this reason it is easy to miss and this document attempts
to provide a list of things to check and under what circumstances this
may apply or not.
manifests at the extreme as the ESX host servers stop responding and
have to be rebooted to come back. Before the reboot the VM's continue
running and the host could respond to pings but it was not responsive
otherwise such as to ssh logins via putty. In less extreme cases the
symptoms may not be as noticeable.
I have seen cases where it was not possible to vmotion or RDP to any VM's running on affected hosts. Which is why I recommend checking before you run into problems.
The way to check is either look at top output (it will be at the top of top if it is really bad) otherwise ps output. For example: ps -auxxwww | grep vem-swiscsi
Check for vem-swiscsi process if unknown ESX hang occurs