VMware Cloud Community
depping
Leadership
Leadership

Clone and roll out template fail with "failed to connect to host"

I have the following problem:

When creating a clone or rolling out a template I always receive the error "failed to connect to host". It doesn't matter if we clone from local storage to san or from local to local or san to san. It never works. The weird thing is that there are hardly any errors to be found in /var/log and copying on the console works fine, even with vmkfstools. I have no problems starting or creating a VM on local or san storage.

The hosts are ESX 3.5, with all 9 patches, and VirtualCenter 2.5.

Duncan

My virtualisation blog:

0 Kudos
14 Replies
depping
Leadership
Leadership

For some weird reason /etc/opt/vmware/vpxa/vpxa.cfg contained a wrong ip in the field hostip. changing this didn't help, but emptying the database with vpxa.exe -b solved it.

Duncan

My virtualisation blog:

0 Kudos
lundpaulsen
Contributor
Contributor

Hello

I have det same config and the same problem..

Where do i run vpxa.exe -b ?

My VC2.5 do not have vpxa.exe

BR

LundPaulsen

0 Kudos
espenam
Contributor
Contributor

I ran into the same problem after changing the IP address for COS on all (two) ESX servers (3.5) in a cluster. Both servers had the wrong (old) "hostIp" in /etc/opt/vmware/vpxa/vpxa.cfg (config file for the VirtualCenter Agent).

Editing the file didn't work. I then simply removed the hosts from the cluster and added them again (one at a time). That solved my problem - cloning worked fine again!

-Espen

0 Kudos
Da_Monkey80
Contributor
Contributor

I had the same issue after I changed the IP of the COS. I also removed the hosts and re-add them back to VC and it worked.

0 Kudos
radman
Enthusiast
Enthusiast

I had the same issue after I changed the IP of the COS. I also removed the hosts and re-add them back to VC and it worked.

I am having the same issue, after changing the IP of the COS. I removed them from Inventory in the VC, then added them back (by browsing the Storage for the Virtual Machine, right-clicking, and selecting "Add to Inventory"), but I still have the same issue. When you say "removed the hosts" do you mean "remove from Inventory" in the VC itself?

0 Kudos
Da_Monkey80
Contributor
Contributor

No, remove the host from VC completly by right-click on the host itself and click remove (not the template). Then right click on the cluster and "Add Host".

Do this to all the hosts that you changed the the IP of the COS.

0 Kudos
Da_Monkey80
Contributor
Contributor

BTW -- you will lose all performance data collected for the host and any guests on the host.

0 Kudos
radman
Enthusiast
Enthusiast

Aha - this did the trick, thanks.

BTW, when I added the host back in, all the VMs came back automatically - great news.

0 Kudos
Da_Monkey80
Contributor
Contributor

Good to hear - Yea meant you would lose the performance metrics for the guests (and hosts)- not lose the guests themselves... For some reason people rely to heavily on that performance metrics and have a heart attack when they see that data gone.

0 Kudos
squirrelking
Enthusiast
Enthusiast

Great, identical problem and solution worked for me too.

Thanks!

---- VCP3 VCP4 VTSP VSP VMware Solution Provider Premier Partner VMware Authorized Consultant [http://blog.lewan.com]
0 Kudos
abeleski
Contributor
Contributor

I have the same issue. When I try and remove the host from VC it tells me that the host must be in maintenance mode before it can be removed. Unfortunately the host has running VM's on it and I do not have capacity on another host within same cluster to vmotion them to.

Is there no way to remove a host from VC and at same time leave the VMs on it keep running? I then want to add it back to VC so I can get rid of the problem i am having due to wrong IP in vpxa.cfg

Thanks

0 Kudos
Da_Monkey80
Contributor
Contributor

Yes you can remove a host from VC without shutting down the guests on it. This sounds harsh, but I do this all the time in my production environment.

  • 1) Disable HA and DRS on the cluster (note any special DRS rules)

  • 2) Right click on the host and select "disconnect" (wait for the task to complete)

  • 3) Right click on the host and select "remove"

  • 4) After the removal is completed you can re add the host back to the cluster.

Hope this helps

-ck

0 Kudos
abeleski
Contributor
Contributor

Ahhh what a good idea. Considering we don't have any complicated DRS rules and we don't have HA enabled this should work for me. I will try it out. Thanks for the quick response.

0 Kudos
smudge1977
Contributor
Contributor

Rather than loose all you performance data this article:

*

This sql should update:

UPDATE .[dbo].[VPX_HOST]*

SET = ‘w.x.y.z'

*WHERE = ‘name of esx host as it is listed in the table'

*

0 Kudos