Virtual Center and ESX Connect issue

[spi200]

All

I have one ESX 3.x Server in a Cluster which VirtualCenter shows as disconnected, when I attempt to connect it jsut says connecting until it times out and says the server is not available or that there is a network issue. The server is available and we do not have a network issue (its on the same segment) - the VM's on the server (about 20) are running fine and I can putty to the console etc.
I have bounced the ESX and Virtual server a couple of times without success. We are running V. 2.5 Virtual center and all servers in the Cluster are 3.x

Any tips would be most helpful


Dave

 

[spi200]

All

Here is some more details on the error I receive.

Error message when attempting to connect ESX server host within VirtualCenter:
Unable to access the Specified host. It either does not exist, the server software is not responding, or there is a network problem.

Information for Virtual centre logs:
The specified network name is no longer available.
Unexpectant error logging into servername.domain.com.au Set connection error code (-2) for host.

I have attempted to connect the ESX server by IP address and also removing and re-addding into Virtual center without success.

Thanks

Dave

 

[TreyJ]

Check your log files. If they are full, the service will stop.

 

[spi200]

HI Treyj

Thanks for the response.

There are a number of logs on both the ESX and Virtual centre, which log do you mean and how do I know if its full or not.
I am gathering the logs for VMWARE support guy.


Regards


Dave

 

[Arisap]

also make sure that there is a directory on the ESX server called /tmp/vmware-root. The VC lays down the VPXA aget on the host. It needs this directory to do so.

 

[spi200]

Hi Arisap


the Vmware-root folder did not exist so I created it, I removed the server from Virtual center and re-added it - Still no joy.

Keep the suggestions comming as VMWARE support is hopeless.


Dave

 

[egodette]

is the server(host) listed in your DNS? If not, addit and it should connect.

 

[stupman]

Make sure that you dedicate enough memory to the service console. You can change the setting in Config/memory. The default value is 272MB, but VMware has already declared that this is way too little.
After you've made the change, you'll have to reboot the ESX in order for the changes to take effect.

 

[spi200]

Thanks guys

Here is an update:

I have entries in my hosts file for Virtual center and the license server, so no DNS to worry about.
The memory was at the default of 272Mb, so I have increased it, but might be a week before I can shut it down.

Vmware support have requested more logs from the server, these guys are slow ot respond.

Keep the suggestion comming.


Dave

 

[Provogeek]

you are speaking of 3.5 right? If it is 3.02 or 3.01 there are other paths to follow. (you only specified 3.x, which depending on point of view could mean any of the three versions).

Even I like to have the entries in my hosts file, but you should still make sure your network DNS has entries for your ESX and VC servers as well as ensuring you have reverse DNS lookup properly configured.

Have you tried to connect straight to the ESX box using the VI client?



=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Brent Schmidt SPOOOOON!!!!!
Senior Network Engineer
Keep IT Simple

http://www.kiscc.com

 

[spi200]

Thanks Provogeek

Sound like you are an old Novell Admin with that name...


The version is 3.02 and was find until our Admin upgraded oure VC from 1.x to 2.02 - he did litte about it and has now left.
The Server was one of three in a cluster, I can still manage the otehr two via VC.
the VI client works fine as well to this server and is how we have been working now for a couple of months - would really like to get it connected again and added back into the cluster so I have some more failover so I can do some patching etc.
I have checked with some pings to FQDN from my workstation & the ESx server and all resolve correctly for the VC, ESX and License server.


Regards

Dave

 

[Provogeek]

Try:

service mgmt-vmware stop

/etc/init.d/vmware-vpxa stop

rpm -e VMware-vpxa-2.0.1-32042 (to delete the agent manually)

service mgmt-vmware restart

Leave host for a few minutes and then reconnect in VC.

=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Brent Schmidt SPOOOOON!!!!!
Senior Network Engineer
Keep IT Simple

http://www.kiscc.com

 

[Provogeek]

btw

you have verified your server is fully patched right? You won't get a VC2.5 server to talk to a 3.02 server with out the 3.02 server being at the latest code.

=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Brent Schmidt SPOOOOON!!!!!
Senior Network Engineer
Keep IT Simple

http://www.kiscc.com

 

[spi200]

Thanks provoGeek

I will try this today.

We have around 25 ESX servers including 2.5, 3.0.1, 3.02 and 3.5 in various stages of patching and all work fine except one.
My version is VMware ESX Server 3.0.2 build-61618 and its one of the later builds that was recently installed - about three months ago. I can patch it on the weekend if you think its not up to scratch, but I have others at build 52797 and they are fine.
I removed it from VC before attempting to connect and then re-added the host. It got to 80% and then just hangs - probably till it times-out.
I seem to recall the previous admin saying something about hanging at 80%, maybe he already tried to do this.

Regards


Dave

 

[spi200]

Thanks to all who contributed.

I got this resolved after weeks of playing around and repeated calls with VMWARE support.
Analysis of the log files showed problems connecting to the NFS storage, even when the VM's were running fine. On the weekend we were upgrading our NFS storage and I noticed the hosts file on the ESX server has no entry for the filer.
I added it in and it connected straight away.

Thanks again

Dave