The problem has to do with SSH checking the known_hosts file. Try adding the
following to /etc/ssh/ssh_config before the "Host *" section:
Host vmguest-* vmhost-* 10.*
This will cause SSH to ignore the known_hosts file for your VMs and VM hosts on
the private network. The host keys change often as computers are reloaded.
Adjust the Host line to match your computer naming convention and private IP
I would also add the following settings to ssh_config to cause things to timeout
faster if a computer is not responsive:
Also, add the following so you don't have to specify the identity file every
time when using SSH via the command line:
Hope this helps,
Terry McGuire wrote:
On 1 Feb 2010, at 0750h, Andy Kurth wrote:
It looks like there are some configuration details which aren't set regarding
your VM and the VM host server it resides on.
Each VM is assigned to a VM host, and each VM host is assigned a "VM Profile".
The error you included indicates that the VM host's profile is not set. To check this:
-Click on "Virtual Hosts". Do you see the host you're using in the list under
"Select a Virtual Host"?
-If yes, select it and click "Configure Host". What do you see next to "VM
-If no, then your host was not correctly set to the vmhostinuse state and
assigned a VM Profile. This will require some DB manipulation. I can provide
instructions if you send some table dumps (listed later).
If your VM host appears to correctly have a VM Profile assigned to it, the
problem could also be the result of your VM not being assigned to a VM host.
To check this:
-Click on "Virtual Hosts". Select the host you are using and then click "Configure
Host". Is the VM you are using is assigned to the host?
Thanks! The above got me further than I was before. I had not created a vmhost by
setting the management node to "vmhostinuse". The related pieces fell into
place after that.
Now, while there are fewer errors in the log, the capture is still failing:
2010-02-01 10:51:25|10975|1:7|image|image.pm:process(161)|calling provisioning
module's capture() subroutine
2010-02-01 10:51:25|10975|1:7|image|vmware.pm:capture(1175)|req=1, res=7: new
2010-02-01 10:51:25|10975|1:7|image|vmware.pm:capture(1176)|req=1, res=7:
2010-02-01 10:51:25|10975|1:7|image|vmware.pm:capture(1177)|req=1, res=7:
2010-02-01 10:51:25|10975|1:7|image|vmware.pm:capture(1178)|req=1, res=7:
SSH command on vmguest-1:
|10975|1:7|image| /usr/bin/ssh -i /etc/vcl/vcl.key -l root -p 22 -x vmguest-1 'chown
root currentimage.txt; chmod 777 currentimage.txt' 2>&1
|10975|1:7|image| ---- WARNING ----
|10975|1:7|image| 2010-02-01 10:51:25|10975|1:7|image|utils.pm:run_ssh_command(6252)|attempt 1/3: failed to execute SSH command on vmguest-1: chown root currentimage.txt; chmod 777 currentimage.txt, exit status: 255, SSH exits with the exit status of the remote command or with 255 if an error occurred, output:
|10975|1:7|image| ssh output (chown root...): Host key verification failed.
|10975|1:7|image| ( 0) utils.pm, notify (line: 737)
|10975|1:7|image| (-1) utils.pm, run_ssh_command (line: 6252)
|10975|1:7|image| (-2) utils.pm, write_currentimage_txt (line: 6500)
|10975|1:7|image| (-3) vmware.pm, capture (line: 1181)
|10975|1:7|image| (-4) image.pm, process (line: 162)
|10975|1:7|image| (-5) vcld, make_new_child (line: 594)
|10975|1:7|image| (-6) vcld, main (line: 341)
...Followed by many more ssh-related warnings, ending with the same failure as
I'm able to ssh to the vm without password from the command line, which
suggests the ssl keys are where they need to be, doesn't it? This feels like
I'm missing just one or two tiny details. Your continued assistance will be