On 09/08/2016 06:08, Matt Maier wrote:
Because we're capable of building systems more complex than we can
understand. So there are always ghost states it can get into that we didn't
prepare for.
I was reading about "crash only" programming a while ago. It like using the
"turn it off and back on again"approach as a part of normal business. Since
all of your systems need to be able to recover from a crash anyway, why
bother programming a graceful shutdown? Just set them up so that they can
pick up where they left off and crash them if anything isn't running
perfectly.
At the conference, the "vulcanbot" continuous integration service was
mentioned several times. This is a microservice I created to facilitate
our development workflow by linking our GitHub repositories to our build
farm.
vulcanbot uses _exactly_ crash-only error handling; whenever something
unexpected happens, it quits (generating informative log messages), and
then gets automatically restarted by systemd. In practice, this works
incredibly well.
Peter
--
Dr Peter Brett <[email protected]>
LiveCode Technical Project Manager
LiveCode 2016 Conference: https://livecode.com/edinburgh-2016/
_______________________________________________
use-livecode mailing list
[email protected]
Please visit this url to subscribe, unsubscribe and manage your subscription
preferences:
http://lists.runrev.com/mailman/listinfo/use-livecode
