goensch Posted May 24 Posted May 24 Hi, I am sorry for causing trouble again. Could you please unsuspend my account goensch on Tommy? Thanks. Regarding what causes the load: I regularly run some Python scripts. Load used to be about 25 CPU and 30 memory (in the graph, I think it is 2500 and 30000). Then, suddenly memory shoots to 100 (I guess the max of the graph). Of course, I have some ideas for incremental incremental improvements, but it is strange the setup worked for more than a year. Anyways, maybe you guys have ideas: I run the scripts as scheduled tasks, using the "fetch website" functionality. Initially, I searched for a way to "directly" call a python, but did not succeed. Is this possible and my it help? My theory is that a task gets stuck. Then, later, the next one is called, gets also stuck.... so memory usage adds up. Is there a possibility to automatically kill the tasks after , say, 2 minutes if they are still running? Thanks a lot, Jochen
wolstech Posted May 24 Posted May 24 Let's see if Krydos has any advice on this. You've been suspended 3 times for this already...
Krydos Posted May 24 Posted May 24 11 hours ago, goensch said: My theory is that a task gets stuck. Then, later, the next one is called, gets also stuck Yep, at the time of your suspension there were 5 readMeter.py running. Just have the new readMeter.py file check for already running instances of itself and kill them if they exist.
JochenG Posted May 24 Posted May 24 Ok, thanks a lot. I think I already found some code to use a pidfile to check if it is running and also to kill.
Recommended Posts