juju-solutions / matrix

Automatic testing of big software deployments under various failure conditions
Other
8 stars 9 forks source link

Added try-catch around glitch action functions. #42

Closed pengale closed 7 years ago

pengale commented 7 years ago

Sometimes, glitch will do something broken like try to reboot a unit that it has just removed.

As a first pass at fixing this, I added a generic catch and log to glitch actions.

@johnsca @bcsaller

johnsca commented 7 years ago

I think the catch you added was in glitch/actions.py but it turns out we're not getting any kind of exception when this fails. Instead, it's just never returning, even when I added a timeout pass-through arg to libjuju.

Glitch plan: http://pastebin.ubuntu.com/23600803/ Matrix log: http://pastebin.ubuntu.com/23600807/

pengale commented 7 years ago

Closing, as this does not address the actual problem of timeouts, and we don't really want to be squashing exceptions if we don't have to.