Tomcat's H2 database(userconf.mv.db)corruption occurred after server shutdown

[laughsmile]

Hello everyone.

Our KNIME Server is working in AWS EC2 , for cost saving , we shutdown EC2 server in 10 pm and restart EC2 server in 9 am automatically.

on Friday of last week, Tomcat used almost all cpu after server started in 9 am, even after knime service restarted.
We found that tomcat’s H2 database(userconf.mv.db)corruption occurred in the server’s shutdown time. Tomcat was executing write operations to H2 database(userconf.mv.db) In the server’s shutdown time.
It said that a database corruption may occur in the H2 database for an abrupt server shutdown.
We solved the issue by using a default userconf.mv.db.

It said that for production environments, the H2 database is not supported and using an external database is recommended.

We want to know how to prevent the recurrence of this problem?
Do we have to build a program to stop knime service before shutting down EC2 server?

Thanks in advance.
Ryu

[thor]

Yes, you should absolutely first shut down KNIME Server before you stop the instance! This is true for every service you run on a server.
The KNIME Server installer comes with systemd script that start and stop KNIME Server automatically during boot and shutdown. See KNIME Server Installation Guide for details.

[laughsmile]

Hello @thor

thank you for your prompt reply.
We are using Windows Service to start and stop KNIME Server automatically during boot and shutdown.

After changing tomcat’s max memory pool from 1500M to 2000M, it seems this problem has stopped recurring.

Best regards
Ryu