Not getting the machine.cfg file after agent install - show "waiting for agent"

[tehv1s1on]

This is a test of Linux desktop with view 7

I have tried many things but that file does not want to show up. The only thing I can see that might be an issue is the linux desktop I try to prepare is on a different domain than the view server, and its subnet is not a normal class C. Ports are open and I can resolve FQDN without the cheat of having it in the /etc/hosts file. Any thoughts?

Works fine an a different linux machine on the same domain and normal class C like the view server.

Thanks for any help.

I have tried to go all the way back to reinstalling vmware tools too.

[tehv1s1on]

Am I going about this wrong?

[zhiminli]

Hi,

Could you please collect the log bundle so we can have a look?

You can generate the bundle via below command.

/usr/lib/vmware/viewabent/bin/dct-debug.sh

BTW, it will also helpful if you can provide some more information about the IP/domain configuration.

Thanks,

Zhimin

[tehv1s1on]

I hate to say this but I can't dump the logs here. I have looked over all the pieces and this is the info I can provide. In the debug log and the error and debug state are.

Error <FileMonitor for /var/vmware/viewagent/MachineId> [UpdateMachinIdhandler]Error watching machine id File file length:2671 exceeds  max 2048  ---> now the exceeds one I also see on the on that is working so not sure if it matters

Dedug<script Stout [linuxUtilities] vmMonitorloginUser: envInfo file don't exist MonitorLoginUser

As for network info that I think might be a problem is the view admin server is on domain x.com class C subnet

Linux vm with agent is special because it has a nist domain of y.com but it's DNS domain is z.com - > so in the log dump it shows in the domain.txt the nist domain(y.com) but its real network for eth0 is the z.com. Pings and communication do flow correctly through DNS between the view server and the linux vm.

[tehv1s1on]

Any thoughts?

[txiong]

Hi ,

"waiting for agent", which will be updated to "unreachable" by Connection Sever after a while, generally means that agent cannot communicate with Connection Sever.

From your description, I guess you're using the managed VM, which is version 7.0.1 (Please correct if I'm wrong).

So I think you need double check:

1. broker has add machine.id option to the VM's vmx file

2. In the Linux VM, you can get the machine.id with command: #vmware-rpctool "machine.id.get"

3. If agent VM has multiple networks (IPs) setup. if yes, you have to specify the Submet option in /etc/vmware/viewagent-custom.conf : Setting Options in Configuration Files on a Linux Desktop

4. "ping <broker FQDN>" can be resolved to IP

5. broker port 4001/4002 are allowed by firewall for agent's JMS communication

[tehv1s1on]

I have done #2 and it pulls what looks like should be in the file that doesn't get generated. There is only one nic so I don't think #3 matters but during testing I added the subnet anyways. #4 resolves. Was promised by networking crew the #5 is good, but I will revisit them.

For #1 I will take a look at the vmx file but I thought since I got a return from the rpctool that I was good.

Thanks for taking time to help. I need to get this thing talking.

[txiong]

what does "during testing I added the subnet anyways" mean?

If the LinuxVM has multiple IPs and not all can reach to View Connection Server, then Subnet option should be used.