Virtual machine node stops responding¶

If one of the control plane VM nodes stops responding, you may need to redeploy it.

Workaround:

1. From the physical node where the target VM is located, get a list of the VM domain IDs and VM names:

   virsh list
   

2. Destroy the target VM (ungraceful powering off of the VM):

   virsh destroy DOMAIN_ID
   

3. Undefine the VM (removes the VM configuration from KVM):

   virsh undefine VM_NAME
   

4. Verify that your physical KVM node has the correct salt-common and salt-minion version:

   apt-cache policy salt-common
   apt-cache policy salt-minion
   

   Note

   If the salt-common and salt-minion versions are not 2015.8, proceed with Install the correct versions of salt-common and salt-minion.

5. Redeploy the VM from the physical node meant to host the VM:

   salt-call state.sls salt.control
   

6. Verify the newly deployed VM is listed in the Salt keys:

   salt-key
   

7. Deploy the Salt states to the node:

   salt 'OST_NAME*' state.sls linux,ntp,openssh,salt
   

8. Deploy service states to the node:

   salt 'HOST_NAME*' state.sls keepalived,haproxy,SPECIFIC_SERVICES
   

   Note

   You may need to log in to the node itself and run the states locally for higher success rates.