Still no luck. Meanwhile, I came across two old posts and appearently the easiest/fastest solution is reboot (even suggested by the support), or restart xinetd service.

Will restart xinetd service affect my VMs?

http://communities.vmware.com/thread/119675

http://communities.vmware.com/message/828264#828264

root@esx1234 root# ps ax | grep D | grep -v grep
PID TTY STAT TIME COMMAND
21604 ? S 0:00 /opt/hp/hpsmh/sbin/hpsmhd -DSSL -f /opt/hp/hpsmh/conf /smhpd.conf
21620 ? S 0:00 /opt/hp/hpsmh/sbin/hpsmhd -DSSL -f /opt/hp/hpsmh/conf /smhpd.conf
root@esx1234 root# ps ax | grep Z | grep -v grep
root@esx1234 root#

----root@esx1234 /--# service mgmt-vmware status
vmware-hostd is stopped

--root@esx1234 /--# service mgmt-vmware start
Starting VMware ESX Server Management services:
VMware ESX Server Host Agent is already running
Availability report startup (background) OK

--root@esx1234 /--# service mgmt-vmware status
vmware-hostd is stopped

I found that the PID files weren't cleaned up after the hostd agent was stopped the last time. Therefore when you try to restart the agent with service mgmt-vmware restart, the script finds what it thinks is a "life" PID for the vmware-hostd daemon, and therefore gives the "VMware ESX Server Host Agent is already running" line.

This is what I had to do to get past this:

If the commands "ps -auxwww |grep -i hostd" or "pidof vmware-hostd" returns no process IDs , you know that the management agent is not running! We therefore need to find the files that reports a false PID.

As the root user, do:

cd \var\run\vmware
ls -l

If any of the following files exists, delete them:
vmware-hostd.PID
watchdog-hostd.PID

Once you are sure those files are deleted, try restarting the management agents again:

service mgmt-vmware restart

-or-

/etc/init.d/mgmt-vmware restart

Hope this helps a bit.

Regards

Rynardt Spies
VCP, vExpert
www.virtualvcp.com