The Graceful Admin
As system administrators, we are called upon to be responsible for a vast quantity of discrete subsystems, each with its own set of operating rules.
When starting and stopping any subsystem, it is the best practice to do so gracefully whenever possible to ensure data integrity.
A Tungsten Cluster is no exception!
In this blog post, we detail the best practices for stopping and starting a Tungsten Cluster on the command line. Please note, you may also perform these operations easily using Tungsten Dashboard GUI (included at no extra charge with your Tungsten subscription).
How To Shut Down A Cluster Gracefully
- Place the cluster into MAINTENANCE mode
- Run `
echo set policy maintenance | cctrl
` - Without this step, all the rest will fail because the Manager will try to put the Replicators back online
https://docs.continuent.com/tungsten-clustering-6.1/operations-policymodes.html
- Run `
- Gracefully offline all Replicators on all nodes
- Do the replicas first, then the primaries
- Issue the `
trepctl -all-services offline
` command
https://docs.continuent.com/tungsten-clustering-6.1/cmdline-tools-trepctl-command-offline.html
- Gracefully stop all running Tungsten cluster processes
- On all nodes, execute: `
stopall
`
https://docs.continuent.com/tungsten-clustering-6.1/cmdline-tools-stopall.html
- On all nodes, execute: `
- You may now proceed with normal OS shutdown or whatever maintenance is needed
How To Start Up A Cluster Gracefully
- Make sure all instances are up and running from the OS perspective
- Start up the Tungsten Cluster processes on the Primaries
- If the cluster services aren't configured to come up automatically at boot time, run the `
startall
` command as the tungsten OS user
https://docs.continuent.com/tungsten-clustering-6.1/cmdline-tools-startall.html - Start-up at boot time is configured using the `
deployall
` script
https://docs.continuent.com/tungsten-clustering-6.1/cmdline-tools-deployall.html
- If the cluster services aren't configured to come up automatically at boot time, run the `
- Repeat for all replicas
- Place the cluster back into policy AUTOMATIC if all nodes are healthy as seen in
cctrl
echo ls | cctrl
echo ‘set policy automatic’ | cctrl
Wrap-Up
In this blog post, we explained the best practices for stopping and starting a Tungsten Cluster. If you’d like to learn more about this or Tungsten Dashboard, please reach out. If you’re a Customer, reach out to Continuent Support via Zendesk; otherwise please submit a Contact Us form.
Comments
Add new comment