We were using ESXi 4.1 u2 standard with Vcenter Standard.
Last week I converted the physical Vcenter server to virtual and all ran fine
then I upgraded the Vcenter 4.1 to 5.1 and then all the esxi hosts one by one and installed the lates patches.
My problem is, after the upgarde, I started seeing some virtual machines are shown in italic font and "inaccessible" next to them. There is no correlation between those vm's. I unregistered and then registered again through data browser and they showed regular. After a while the same problem with other vm's and/or same vm's. Then I unregistered again all of them, deleted all kind of .nvram .lck .vswp files that were as I thought may be those files are disturbing somehow and then registered again. Today I started seeing couple of more machines that are "inaccessible".
The strange thing is, I see the problem only on powered off machines. Powered on vm's go on running without problem.
I don't understand what is causing this problem.
Anyone can help ?
Thanks,
Haim Beyhan
IT Manager
1. can you check the storage status at the time the VM went inaccessible. The datastore might have disconnected for a while and came back.
2. Instead of re-registering the VMs you can just re-load it.
this command from CLI will give you the Name and VMID of the VMs which are inaccessible.
vim-cmd vmsvc/getallvms | grep skip
Skipping invalid VM '118'
this will reload the VM which is not accessible
vim-cmd vmsvc/reload [NUMBER]
eg: vim-cmd vmsvc/reload 118
I checked the vmkernel.log and I saw that the ESXi lost connection to datastore.
After some research, I found out a problem with ESXi 5.X with Netapp storage in which I have to change an Advanced Setting.
I did now this setting and rebooted the server. I hope this will fix it.
Regarding the commandline for loading inaccessible machines, it worked.
Thanks,
Haim Beyhan
good luck.