This problem was the subject of a recent Cheetah Test Run Support Forum discussion. A few suggestions on how to avoid this problem were made..
Stopping the DB Scheduler
Shutting down the DB Scheduler means missing out on some very useful features but it is a reasonable temporary diagnostic strategy if you suspect it may be causing problems or if oncheck -pe shows the sysadmin database is filling up the root dbspace.
The recommended way to shut down the Scheduler is to use the task() routine..
database sysadmin;execute function task("scheduler shutdown");close database;Another method mentioned in the forum discussion was to create an empty file $INFORMIXDIR/etc/sysadmin/stop before starting IDS, which will prevent the Scheduler from starting. This method is not documented and not guaranteed to work in future versions.
Modifying the Scheduler to limit space consumption
Better than disabling the Scheduler altogether is to update the ph_task table of the sysadmin database to limit the amount of data that is collected. This example tells the Scheduler to delete data older than 4 times the collection frequency:
database sysadmin;update ph_task set tk_delete = (4*tk_frequency) where tk_type ='SENSOR';
Mirav's response to the forum question pointed out the relevant section of the 11.10 Admin Guide: Setting up Tasks which discusses setting up and monitoring Scheduler tasks.
On the subject of the sysadmin database - John Miller's Chat with the Labs is at 9am Pacific (5pm GMT) tomorrow (Wednesday Feb 28 2007) entitled Cheetah - Expanding the Administration Free Zone. Having had a sneak preview I can say it's definitely not to be missed.