Every now and then the damus relay just dies and stops accepting posts. Through much debugging, expert analysis, tapping into my years of experience in maintaining software and getting to the bottom of complex computer science issues, my final solution is to just kill the process every hour. Hit that follow button for more state of the art #devops techniques.
Discussion
True and tried method among #devops , can confirm! 🐶🐾😂🫂
The good old turn it off and back on again
When it works it works! Nothing wrong with that!
When I can't figure something out, I just turn it off, unplug the electricity for 2 weeks and then I'll try it again. #likeapro
„Ein Reboot tut immer gut“
how we say in German
I’d say regular reboots are one of the main success facors of #heroku.
I know of a multi-million dollar Enterprise software company who did exactly that, and shipped a maintenance release which only restarted the process regularly under the hood
Of course you couldn't find any of that in the release notes.
Ah the all powerful “unplug it and plug it back in” method. Genius.
I’ve been running roosofts relay for about two months with no hiccups. It doesn’t store any data yet and is very basic but for a high availability and battle tested run time that easily scales, elixir is the way to go. It could easily be extended and the code is very clean and well documented and test driven.
so, turning it off and on again
Be strong be proud of your work 📝
My PC’s Xorg logfiles never seem to prune right so I added an hourly ‘echo "" > logfile’ to my crontab
Where there's a #[1] there's a way. Ingenious stuff.
Have you tried hitting it with a hammer?
I love it, sometimes the best solution is the simplest
Hahahaha