Using HA Force Over to Reduce Downtime when Updating Prod

[Joe Morley]

We are updating our Cloverleaf system from 6.1.0 to 6.1.1 and I would like to confirm that our cut over plan will be a safe and efficient for updating our system.

We have the High Availability functionality in place using two production servers, prod1, prod2. To limit downtime while updating prod we would like use the following steps:

1. Force a fail over from prod1 to prod2, leaving the prod1 server powered up but not processing.

2. Update Cloverleaf on prod1.

3. Force a fail over on prod2, making prod 1 responsible for processing all connections.

4. Update Cloverleaf on prod2

5. Allow prod2 connections to come back up on prod2

Does anyone see cause for concern using the steps above to update our production servers?

Has anyone performed a system update using the technique above? If so could you please leave some feed back on how it worked for you?

Thank you very much for any details that can be provided!

Topic

[Robert Kersemakers]

Hi Joe,

No concerns, just a different approach:

1. Update prod2.

2. Fail over from prod1 to prod2.

3. If all runs smoothly: leave all running on prod2. If there are problems, you can always go back to prod1.

4. After a certain period of running smoothly on prod2, update prod1.

This way you will only have one period of downtime. Assuming that prod1 and prod2 are physically equal.

Zuyderland Medisch Centrum; Heerlen/Sittard; The Netherlands

[David Barr]

You should also stop all of your inbound threads and try to make sure all of the messages are gone from the recovery DB before you fail over.

[Author]

Replies