DELETE hangs on one Galera Cluster node (10.2), on others nodes fine

Hello, I have a problem when executing a complex DELETE (with subqueries) of about 200,000 records in a Galera Cluster of 3 nodes (10.2.36 version).

On one of the nodes, the query hangs (it doesn't finish, I have to kill it after several hours) and on the rest nodes it lasts just a couple of minutes.

It's been running every month for a couple of years and had never been longer than 3 minutes, but now it just won't finish.

I've run an analyze on the table, reorganized several indexes, and even run an alter table rebuild, but nothing has changed.

Database is not very big, I have thought about removing that node and adding it again but I would like to know the reason for the problem.

Any advice?

Thanks in advance,
Jon.