Failover Cluster

In the event that you have critical uptime needs, you can provision an addition ScienceOps cluster to stand side-by-side with your primary production instance. The failover instance will be a mirror of the primary ScienceOps cluster and will be on standby in the event of an issue with the primary cluster. Requests to the server can be made the same credentials, routes, etc. without experiencing any interruption in service.

Recommendations

In a production environment, it's strongly suggested that you use static IP addresses for servers in your failover cluster. This makes configuring the failover switch (and potential updates) easier and leads to less confusion down the road.

If retaining prediction data (API request input/output data) isn't mandatory, it's recommended that you setup the failover cluster without the ScienceOps Prediction Logging feature turned on. This reduces the potential breakpoints for the failover instance without significantly compromising its functionality.

It's still recommended that you save your point-in-time backup files in a remote location. While having a hot failover configured will prevent from long-term outages, having ScienceOps backup is still encouraged as a best-practice.

How To

Initial Setup 1. Provision hardware for "primary" ScienceOps instance, perform installation, and configure users. 2. SSH into the master node. 3. Create a backup of the primary ScienceOps instance by running scienceops-backup. See Backups for more details. 4. Provision hardware for failover ScienceOps instance and perform installation. 5. The backup file from your primary ScienceOps using backup file from Step #3. See Backups for more details.

Ongoing Updates 1. Setup a recurring job at the interval of your choosing (hourly, daily, etc.) to that runs scienceops-backup on the primary server and transfers the output file to a remote location. 2. Setup a recurring job on the backup server that runs the scienceops-restore --checkpoint {filename} on the backup file output by Step #1. This will "catch-up" your backup server to the point in time backup created by Step #1.

results matching ""

    No results matching ""