Closed cyberbit closed 1 year ago
I think it makes the most sense to just catch/log the error and wait until the next go round. The delay is negligible (sixty seconds, assuming the outage is momentary.)
I figured failing outright probably isn't intended.
Yeeeah, I'm a little too used to "just HCF and let systemd
restart the thing."
let systemd restart the thing
Ha, yeah this was my plan for the Docker container regardless. Long live systemd
!
(Forgot to trap errors from another request function - sorry if you already built the previous commit!)
While running satpaper, if the connection experiences a momentary outage (either on client or SLIDER end) while tiles are being collected, the program loop crashes. Here's an excerpt of the log when that occurs:
The exact message and segment of code varies depending on what stage of the loop is running. Maybe a gated "retry x times" flag with exponential backoff would be good here? If retrying each tile isn't feasible, then maybe logging the error was caught and waiting for next schedule to retry. I figured failing outright probably isn't intended.