I think this might depend on which node(s) you lose If you have 2 fast nodes and one far/slow node, you will commit acknowledgement back once the first node acknowledges the write. So if you lose two near/fast nodes, your 3rd node could actually be behind and you will lose transactions. Are you saying that you observed transaction loss in case of single node failure?
On Tue, 17 Mar 2026 at 14:37, Kyle Kingsbury via discuss < [email protected]> wrote: > Good morning everyone, > > I've been working on this off and on for a few months, and thought it > might be of interest to the list: MariaDB with Galera Cluster seems to > lose committed updates even with innodb_flush_log_at_trx_commit = 1, > perhaps due to a combination of network partitions and process kills. It > also allows Lost Update (a violation of Snapshot Isolation) and Stale > Read (which violates claims of immediate, synchronous replication) in > healthy clusters. > > A warm thank you to Teemu Ollakka for the tips on log flushing back in > January. :-) > > https://jepsen.io/analyses/mariadb-galera-cluster-12.1.2 > > Yours truly, > > --Kyle > _______________________________________________ > discuss mailing list -- [email protected] > To unsubscribe send an email to [email protected] > -- Gordan Bobic Database Specialist, Shattered Silicon Ltd. https://shatteredsilicon.net Follow us: LinkedIn: https://www.linkedin.com/company/shatteredsilicon X: https://x.com/ssiliconbg
_______________________________________________ discuss mailing list -- [email protected] To unsubscribe send an email to [email protected]
