High CPU load every 30-40 sec
I have implemented 3 nodes Galera cluster successfully but I can observe high load CPU on main node (I begun create cluster from it) every 30-40 sec. Also there is no big IO on three nodes and no any network latency. I have configuration: 2 Ubuntu server 16 and one Ubuntu server 18. Also I use MariaDB 10.2 and Galera 3.
Have you ever had situation like this and what I need to investigate to determine origin of problem?
Thanks in advance.
Unfortunately, I haven't really seen such behaviour. Having said that, I would:
1. Observe 'top' to see which process is responsible for the load.
2. If that's MariaDB, I'd observe process list/slow query log/performance schema to see if there are any queries that are executed
3. Performance schema may also give you more insight into what is happening in the system.Comment actions
Thank you for your answer. I can tell you about first point:
1. This process is /usr/sbin/mysqld --basedir=/usr --wsrep_start_position=9ed07adf-9e62-11ea-87ea-732d4886bce3:150426
all of three database servers work properly when they are not in Galera cluster.
Maybe now you can help determine the cause of the issue?
In that case you have to try to understand what is really happening in your database cluster. Galera itself does use some CPU for the intra-cluster communication but it should not result in high load. Most likely there is some sort of a process that's executed on a regular basis and it is responsible for the load that you see.
Please sign in to leave a comment.