You are viewing the RapidMiner Server documentation for version 9.0 - Check here for latest version
Limitations
Having a cold swap failover take place for RapidMiner Server comes with a few limitations.
Running and waiting processes
- Processes which are running while the primary server goes down will fail. They will at a certain point be flagged as such by the backup server. Results or logs of those processes will not be available.
- Scheduled processes whose execution have not started before the server crash will be executed by a backup server after the failover.
- Recurring scheduled processes that missed one or more execution due to the downtime will be executed only once by a backup server immediately after the server startup and then are scheduled to the next appropriate time.
Web UI Sessions
Web UI session are valid within one RapidMiner Server instance. They are not persisted and cannot be replicated as in case of cold swap failover only one instance is running at a time. This means that in case of failover active sessions will be lost and any unsaved work and unsubmitted changes can't be recovered. Users who are logged into the web interface of the server will experience that the UI is unavailable during the downtime. When a backup server becomes ready to process the web requests users will be redirected to the welcome page, have to log in again and redo the unsaved work.