If you need to deploy a topology to a node that has been
used for a previous topology deployment, ensure that the node is in
a clean state and that any corresponding node data on the deployment
server has been removed.
About this task
To complete the cleanup of the node, do the following steps
to prepare this node and the deployment server.
Procedure
- Revert your node system where a topology was previously
deployed to a snapshot or backup version of the node. The node must
be in the state that it was in prior to when the topology was first
deployed.
- On the deployment server, you must delete the data that
is stored about a previously deployed node. The deployment server
maintains state and other data about a node system in the deployment
server database. If you are redeploying, or if you replace a node
with the same fully qualified domain name, run the following commands
on the deployment server to remove the data that has been stored about
the node:
$ knife node delete node.fqdn -y
$ knife client delete node.fqdn -y
These
commands reset the node to a clean state for the Chef server.
- Depending on how your node is configured, you might also
need to delete the SSH fingerprint on your deployment server. The
deployment knife commands use SSH to communicate with the node system
to deploy OpenStack.
Use one of the following methods to delete SSH fingerprints from the .ssh/known_hosts file
on the deployment server:
- Run the following command:
ssh-keygen -R node fqdn
- Edit the .ssh/known_hosts file in a text
editor and delete the line that contains your node system.