Instance Missing on Hypervisor

Problem

• Instance goes into ERROR state after failing to start.
• Running openstack server show [uuid] command shows an instance fault such as the following.

Error: Instance [UUID] could not be found on hypervisor.

Environment

• Platform9 Managed OpenStack - All Versions
• Nova

Causes

1. Instance failed to migrate.
2. Disk is full on hypervisor.

Resolution

Instance Failed to Migrate

1. Determine on which host the instance resides on.

# openstack server show [UUID]

2. Check if the instance is running on the host obtained from the previous output.

# virsh list --all | grep [UUID]

3. Confirm that the instance XML file still exists on the hypervisor.

# ls -l /etc/libvirt/qemu/[UUID].xml# ls -l /opt/pf9/data/instances/[UUID]/cfg-[UUID].xml

4. If the XML file resides on the host, redefine the instance using virsh.

# virsh define /etc/libvirt/qemu/[UUID].xml

5. Reset the state of the instance.

# openstack server set --state active [UUID]

Disk Full

1. Identify which partition is full.

# df -h

2. Inspect which folder(s) or file(s) are consuming the most amount of disk space.

# du -xh [partition-path] | sort -h | tail

3. At your discretion, optionally backup and/or remove or truncate the offending large files which are consuming the most space.

Note: If you are uncertain if a particular file may be removed or truncated belonging to PF9, please contact Support.