Several rarely-used CloudBees Flow operations are not supported in clustered mode. For any these operations, you must restart the server in single-server mode:
Changing the database configuration This operation can be done in the web UI using Administration > Database Configuration or with
ectool. After performing this operation, you must rerun ZKConfigTool to upload the changed configuration from the local
database.propertiesfile to ZooKeeper before switching back to clustered mode.
Doing a full import using
Doing a full export using the safe mode with either of these options:
To run a CloudBees Flow machine in single-server mode:
Identify the machine you need to work on and shut down the other machines in the CloudBees Flow cluster.
Verify that the
commander.propertiesconfiguration files on the machine you are working on are up to date and match those that were uploaded to ZooKeeper.
wrapper.conffile of the machine that you are working on. Ensure that the number
261is unique within the file. If not, change the number to a unique ID. This line prevents server mismatch errors after the switch from multi-node cluster mode to single-node standalone mode.
Switch the remaining CloudBees Flow machine to single-server mode by running the following command:
Complete your work on the CloudBees Flow machine.
If you changed the database configuration, use ZKConfigTool to upload the updated
database.propertiesfile to ZooKeeper.
See Uploading Configuration Files to ZooKeeper for more information.
If you used the
eccerttool (for added trusted agents, revoked certificates, and so on), use ZKConfigTool to upload the updated keystore file and/or the
conf/securityfolder to ZooKeeper.
If you updated the
commander.propertiesfile, use ZKConfigTool to upload it to ZooKeeper.
Switch the remaining CloudBees Flow machine back to clustered mode by entering:
Restart the other nodes in the CloudBees Flow cluster.