Issue metadata
Sign in to add a comment
|
mash: Don't automatically restart ws or ash processes if they crash |
||||||||||||||||||||||||
Issue descriptionWe've had some problems with our hardware autotest where ash gets into a crash loop and fills the disk with core files. On-device there's a session_manager process responsible for restarting chrome when it crashes. Medium-term we should bring down everything (including the process running service manager) if ws or ash crashes. In the meantime it would help to not auto-restart the ash service. Is there a config for service manager that allows that?
,
Jan 13 2017
Actually, this is a case where the service (not session) manager, browser and renderers are all still running, but ash has crashed. I think it's the service manager (inside the chrome binary) that restarts ash. I see logging like this: ui.20170112-141554:[32314:32314:0112/142148.221419:263521772758:ERROR:session.cc(22)] Restarting service: ash I think ash is a "restartable service"?
,
Jan 13 2017
(Do we even need mash_session anymore?)
,
Jan 13 2017
,
Feb 26 2018
|
|||||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||||
Comment 1 Deleted