The Ultimate Guide to High Availability with MariaDB

deadlock in Galera cluster

Hi I have been testing MariaDB with Galera cluster for distributed locking. I followed the post from https://severalnines.com/blog/avoiding-deadlocks-galera-set-haproxy-single-node-writes-and-multi-node-reads

I can reproduce the deadlock at will. But interestingly if I run the script in parallel on two nodes only there there is locking. The minute I start the script on the third node bingo deadlock appear.

Hmm I wonder if some settings for Galera can explain that? wsre_thread_count ?

Regards, Steven

Answer Answered by Jeff Dyke in this comment.

I wish i knew when this was written. I am running Galera 25.3.20(r3703) with Maria 10.1.25 in the exact same configuration as the site (3 servers round robin) and can not reproduce this error, unless - like OP, i run it in parallel on 3 machines, two does not do the trick.

Did you have any luck with wsre_thread_count. Moving to a writable failover node in haproxy is really simple for my setup, but if there was another solution, that would be great!

Jeff

Comments

Comments loading...
Content reproduced on this site is the property of its respective owners, and this content is not reviewed in advance by MariaDB. The views, information and opinions expressed by this content do not necessarily represent those of MariaDB or any other party.