HA – Automated Cluster Failovers

Tagged: cluster

This topic has 4 replies, 4 voices, and was last updated 3 years, 5 months ago by Peter Heggie.

Creator

Topic
April 12, 2021 at 5:46 pm #118747

Michael Brande
Participant

Has anyone here done a full failover (using HA) successfully with no manual intervention?

If so, about how long does that process occur?

Any issues with recovery database corruption when doing this?
Creator

Topic

Viewing 2 reply threads

Author

Replies
- April 13, 2021 at 9:05 am #118749
  
  Jay Hammond
  Participant
  
  If by manual intervention you mean at the Cloverleaf level, we have. We’re on AIX 7.1.5 and Cloverleaf 6.2.2, and have application scripts that we point the HA software to that shuts down Cloverleaf cleanly and then starts it back up when the failover is complete. It takes from 2-5 minutes for the failover to finish, and then 10-15 minutes for the Cloverleaf sites/processes/threads to all start back up (10-15 sites, 500-600 threads depending on test or prod environment). We’ve had zero issues with DB corruption when doing this.
  
  Jay
  - April 15, 2021 at 2:52 pm #118755
    
    Michael Brande
    Participant
    
    Thank you Jay!
- April 15, 2021 at 4:17 pm #118756
  
  Devika Agarwal
  Participant
  
  If you are on CIS 19, the scripts are available to you. Let me know if you’d like more details.
  
  Devika
- April 19, 2021 at 5:52 am #118761
  
  Peter Heggie
  Participant
  
  We have never had a system-level event where HA automatically failed over. But we have many times used the HA console on AIX to initiate a failover and it has completed successfully without further intervention. Total time from initiation to threads up and running on the backup server was about three minutes – three shared file systems, network adapters w/virtual addresses, and nine Cloverleaf sites with 100+ threads.
  
  Peter Heggie
Author

Replies

Viewing 2 reply threads

You must be logged in to reply to this topic.