Issue: Unable to connect to NFS datastore and vms inaccessible:-
error 8/21/2015 8:41:02 AM Cannot relocate virtual machine 'amanda client' in Ephox Brisbane
error 8/21/2015 8:37:29 AM Host hostname in Ephox Brisbane is not responding
error: unable to access virtual machine configuration: unable to access file
error; unable to access virtual disk 'Hard Disk 1' from the host either host disconnected from the datastore or has insufficient privilege: unable to access file [stratvms]
Infrastructure: ESXi vpshere 6.0 (vsphere essentials)
Issue first noticed: within last 24 hours
VMware tools: Current
Number vms in cluster:286
Problem vms: 12 in total but only the 5 that are connected to one particular host of our 3 server hosts
after attempting to migrate some vms, to stratvms freebsd 10.2 nfs datastore from different host/datastore the migration of the vm stuck at 36%
error:vmdk file on vatvms. /export/vatvms ~ has mismatching owner/perms due to maproot=0 but this may not our issue. MTU was misconfigured as you have to set at both the vmkernel and the vswitch for MTU 9000. All three hosts were misconfigured though.
zpool iostat at destination showed no disk writing taking place
It was noticed the vsphere client was hanging and received unresponsive error. this morning restart management agent attempted:- service vpxa restart. hostd service runningOK. this is the service agents web client talks to. on milkshake: /etc/init.d/vpxa restart. this service seemed stuck. kill -9 34776 (vpxa) didn't fix. restarting esxcfg. Server rebooted from console screen. After restart all vms came back up. This still hasn't resolved as the hosts seem to be flapping with thier connections to stratvms. I repatched the datastore to a different switch port to make sure it wasnt affecting. host can't talk to datastore so did a right-click, unmount on stratvms, then selected host.
Then did right-click, mount on additional datastores, then added host again. Nothing we do seems to make these 5 hosts accessible again.
Any thoughts appreciated