I'm having problems with the 2.5 Modeller xas instance locking up frequently. When this happens the browser is unresponsive, or returns no data, although the modeller itself can still open microflows/forms and make changes. Restarting the session fails and you have to force a shutdown. This is 220.127.116.11 with built-in database. Occasionally I get error messages in the console while this is happening - like 20 connections to DB in use and cannot obtain a connection... or a scheduled event has failed with an exception. I have noticed that if I look at Scheduled actions in a browser, there are more than 60 in a status of Running. I have disabled the scheduled events in the modeller, but this makes no difference, they still appear. How do I recover from this situation? Is it possible to cancel these running scheduled events? Is it possible to clean-up (delete) the history of scheduled events which has now grown to about 10,000. What might cause a scheduled event to not complete properly?
It sounds like you are running a lot of concurrent scheduled events. The default setting for database connections between the Mendix Runtime and the database is 20. If all these connections are in use you cannot get a new connection and the request from the client cannot be handled. You will see log messages that all connections are in use.
You can tune your database connection settings via the Service program as described on this page.
Johan den Haan
There are only 2 scheduled events involved - running about every 3 minutes. I have 63 instances of the scheduled events in a running status - some dating back 3 weeks or more. Neither of the scheduled events is processing a large number of data records.
Restarting the modeller does not change the status of these running events, nor does disabling the scheduled events. Note that this is using the built-in database/modeller only.
Having scheduled events in the running status can mean two things:
They are actually running, or
The status hasn't been updated because of (for example) stopping the server before the scheduled event was finished.
To fix the last issue you can delete the status records. It can happen when you are restarting a lot because you are developing an application.