I've had a persistent problem with a truly trivial little SOAPpy server silently failing with no exceptions, no messages, nothing. So, needing to get this critical feature working ASAP, I created a bog-simple watching script that checked for that SOAPpy server and restarted it when necessary, on the theory that whatever was killing the SOAPpy server wouldn't kill such a bog-simple script. And that script silently failed. So I created a cron job that ran every 1 minute, doing nothing but checking for the server and restarting it if it failed.
That was where I stood last night.
This morning I woke to messages telling me the server (a crucial one) is dead because of hundreds of these processes. The operating system was hanging onto each and every cron process, apparently because they had an open pipe. Much stressy rewriting to add checking to the script to prevent there ever being more than one instance.
Worked fine for a few hours, then I succumbed to the (biological) bug I've been battling for the last few days for a few hours. Woke up (again, getting out of bed causes the problem) to find the server frozen again, this time without a way to reboot it.
I still have no idea why the SOAPpy server is exiting, but now I've got a whole new crop of headaches on top of the ones from the flu or whatever this is.
In other news, grams tells me that apparently her senior's complex is battling a flu outbreak and has shut down all outside visitations as a result. She hasn't caught anything yet. My fever (which wasn't that high to begin with) has come down, so it seems likely I'll survive if the sysadmins don't kill me. I'm expecting this is some collosally stupid thing such as not getting the code updated to the new version. However, I'm thinking I'll ratchet down the cron frequency to once every hour until I'm sure it's working (assuming I can get back onto the box).
Pingbacks are closed.