meltin.net

This is a report on the status of CTDB, similar to that presented at recent sambaXP conferences. As usual, this presentation will look back and summarise progress since the most recent presentation in 2020. It will also look forward and attempt to present a realistic path for further development.

The biggest recent change arrived in Samba 4.16. CTDB's recovery lock is now a cluster lock and, when enabled, a race for this lock is used in place of a traditional election. This avoids problems where an election would result in a new leader but this leader would be unable to take the lock. Reasons for this include races and cluster filesystem latency.

In the past we have presented grand plans, designs and frameworks. This year we will step back a bit and consider what is needed to realise a shiny new, maintainable CTDB.