We have three ESX 3.5U2 servers in our cluster. Two of them have about 10-15 hostd processes, consuming, together, 90%+ of the console CPU at all times. This makes any VM operation painfully slow (powering on VMs, reconfiguring etc). Here is an extract of the hostd.log:
2009-06-29 10:52:21.425 'vm:/vmfs/volumes/7022384b-332c5a3d/archived vms/wksTest1/wksTest1.vmx' 18987952 info State Transition (VM_STATE_RECONFIGURING -> VM_STATE_ON) 2009-06-29 10:52:21.439 'TaskManager' 18987952 info Task Completed : haTask-pool7-vim.ResourcePool.updateChildResourceConfiguration-4097 2009-06-29 10:52:21.443 'TaskManager' 104192944 info Task Created : haTask-pool1-vim.ResourcePool.updateChildResourceConfiguration-4098 2009-06-29 10:52:21.487 'vm:/vmfs/volumes/47a2df2f-637f3204-c8f1-0016355a7dee/svr1/svr1.vmx' 104192944 info State Transition (VM_STATE_ON -> VM_STATE_RECONFIGURING) 2009-06-29 10:52:21.651 'vm:/vmfs/volumes/47a2df2f-637f3204-c8f1-0016355a7dee/svr1/svr1.vmx' 104192944 info State Transition (VM_STATE_RECONFIGURING -> VM_STATE_ON) 2009-06-29 10:52:21.685 'TaskManager' 104192944 info Task Completed : haTask-pool1-vim.ResourcePool.updateChildResourceConfiguration-4098 2009-06-29 10:52:21.699 'TaskManager' 19254192 info Task Created : haTask-pool3-vim.ResourcePool.updateChildResourceConfiguration-4099 2009-06-29 10:52:21.699 'vm:/vmfs/volumes/47a2df2f-637f3204-c8f1-0016355a7dee/svr8/svr8.vmx' 19254192 info State Transition (VM_STATE_ON -> VM_STATE_RECONFIGURING) 2009-06-29 10:52:21.925 'vm:/vmfs/volumes/47a2df2f-637f3204-c8f1-0016355a7dee/svr8/svr8.vmx' 19254192 info State Transition (VM_STATE_RECONFIGURING -> VM_STATE_ON) 2009-06-29 10:52:21.944 'TaskManager' 19254192 info Task Completed : haTask-pool3-vim.ResourcePool.updateChildResourceConfiguration-4099
This is performed constantly, multiple times per second.
I've tried a service mgmt-vmware restart, but they respawn like the devil, within seconds...
Try to restar the ESXi services.
Are you using resource pools?
So I smashed the small nail that was the problem with a sledgehammer and upgraded U2 to U4, so all hosts had a reboot. I've just now noticed that the threads in ESXi, I'm sure I posted in ESX.... this is not the i product.
So unfortunately this isn't answerable now Sorry, but I needed a result asap!
Try only to restart the mgmt-vmware service with:
service mgmt-vmware restart
Are you using DRS with resource pools?
Yes, I said in my first post that I'd already tried that - the processes disappeard, then re-spawned within 1 minute.
Yes, we have HA/DRS & Resource pools.
As I said in my previous post, this is now essentially solved (unfortunately without finding the issue), but I'm happy to answer any questions people have about the configuration.