Host fails to reconnect to VC.

[Pratt]

I have an ESX Host that is at version 3.5 and VC is 2.5. The Host was showing as disconnected in VC this morning and when I attempted to reconnect it the job has shown 'In Progress' for over 20 minutes and then failed to reconnect.

I have run the service mgmt-vmware restart command but that did not seem to help.

Would anyone have any suggestions as to why this may be happening and how I could get it to reconnect? Thank you.

[larstr]

Try restarting your VirtualCenter service:

net stop "VMware VirtualCenter Server"

net start "VMware VirtualCenter Server"

Lars

[Pratt]

I stopped and started the service on our VC Server and restarted the VIC. The Host still showed 'Disconnected' so I right-clicked the Host and tried to 'Connect' but it appears to be doing the samething, i.e. showing 'In Progress' for a long time and I think it will eventually fail.

[slayedbylucifer]

Follow these steps in hte sequence given, they always work for me, hope it will be helpful.

1) Log in to the ESX host via SSH/PuTTy as root

2) Remove any exisitng VC agent from the esx host:

Run follwing commands to get the names of the packages to be removed:

1. rpm -qa | grep vpxa

2. rpm -qa | grep aam <<<< this package may/may not be installed

3) Remove above packages by running following command:

1. rpm -e <package_name>

4) Re-run commands mentioned in step 2 to ensure that those packages have been uninstalled.

5) Now on the VC server, log in as adinistrator and go to the virtual center installation directory. It usually is C:\program files\vmware\vmware infrastructure\virtual center

6) Here, you should find a folder called "upgrade", change to that folder.

7) Here look for a XML file called "bundleversion.xml", open it.

😎 Now, say for example we want to install the VC agent for ESX 3.0.2 host. So look into this XML file for this verstion of esx. it would most probably be written 3.0.0+ in the file.

9) Now note down the bundle id written right above that version, say 6.

10) Once you have the bundle id, close the XML file and look for the corresponding bundle id file in the "upgrade" folder.

11) Then move this file with the corresponding bundle id to the esx server.

You may use WINScp for this purpose. You do not have to copy the .sig (signature) file to the esx server.

12) This file which you have copied to the esx server is nothng but a shell script. Make it executable by running following command:

1. chmod 755 <NAME_OF_THE_BUNDLE>

13) Run this script using following command:

1. ./<NAME_OF_THE_BUNDLE>

This would install the VPXA daemon and will start it.

14) Run the following commands to restart the services:

1. service mgmt-vmware restart

2. service vmware-vpxa restart

15) Now log in to the VC server as administrator and then add that ESX host to the invetory and it should work.

[espi3030]

Has anything changed on the host? Can you log onto the host, then try to ping the VC server. Also check the /etc/hosts file on the host with issues. Can you remove the host from VC and try to re add it?

[dmismash]

Has anything changed on the switch ports? Is this sitting behind firewall rules? Is the server disconnected in VC in a datacenter or cluster?

We had a problem at a client's site where they made a change on the fw and the hosts would be disconnected if they were added into a cluster. The ports for HA were not open

[adolopo]

Yah....are you able to get to the VC management ports (902, etc)?

[Pratt]

The problem was I had three (3) process running on the ESX Host that had the processor at 100%. I was not able to kill the two (2) vmware-hostd processes and the one (1) vxpa process which was defunct. I had to down my guest systems and reboot the ESX Host which solved the problem. I think the server was just to busy to process the 'Connect' request. Thank you for all the suggestions.