pintail
Posts: 32
Joined: Wed Aug 27, 2014 6:58 pm

System Execute Stops Working after a while in PROD, still works in DEV

Postby pintail » Thu May 18, 2017 3:10 am

Hi,

Version 4.1

We're finding an issue where a system execute call to a python script stops working after a while in production (seems random). The same script will work in DEV. The system execute is called repeatedly by a jobflow which passes a parameter to be used in the system execute component which is used for an API data call to retrieve external data.

I'm wondering if there is something that needs to be done to shut things down after each call? Just guessing here, but just strange that it will work for a while, then complete stop with a URL call error from python, but the same call works in dev, so seems server specific to me and something related to a build of up running repeated and often sys execute graphs.

any thoughts?

thanks!

pintail
Posts: 32
Joined: Wed Aug 27, 2014 6:58 pm

Re: System Execute Stops Working after a while in PROD, still works in DEV

Postby pintail » Thu May 18, 2017 6:06 pm

looks like it's a socket issue on the webserver. We're temporarily just restarting the service to clear things out night, but would be good to see if anyone has any other ideas about the issue. I do see the python processes starting and going away in the monitoring window on linux, so it looks like the process shuts down, not sure why the socket isn't cleared out though still...