840712 - deviceManagerSUT.py:killProcess should retry

Status: RESOLVED FIXED

(Testing :: Mozbase, defect)

Description

[Bob Clary [:bc] (inactive)]

Attachment: patch

devicemanagerSUT.killProcess will abort if it fails to kill the process. This is a problem since the inability to kill processes sometimes at least appears to be transient. I've seen this often in autophone.

We should try to kill the process up to the limit devicemanagerSUT.retryLimit.

I've tested this on autophone and it appears to work fine.

Comment 1

[Joel Maher ( :jmaher ) (UTC -8)]

Comment on attachment 713100 [details] [diff] [review]
patch

Review of attachment 713100 [details] [diff] [review]:
-----------------------------------------------------------------

::: mozdevice/mozdevice/devicemanagerSUT.py
@@ +570,5 @@
> +            except DMError, err:
> +                retries +=1
> +                print ("WARNING: try %d of %d failed to kill %s" %
> +                       (retries, self.retryLimit, appname))
> +                print err

there is a danger here of the 'err' text geting caught in the tbpl log parsing and reporting an error.  I don't recall any bugs on file in recent times where we have had this in tegra/panda automation, so this might be ok for now.

Comment 2

[Bob Clary [:bc] (inactive)]

(In reply to Joel Maher (:jmaher) from comment #1)

> 
> there is a danger here of the 'err' text geting caught in the tbpl log
> parsing and reporting an error.  I don't recall any bugs on file in recent
> times where we have had this in tegra/panda automation, so this might be ok
> for now.

The other code which prints err qualifies it with if self.debug >=4:, that might explain why you haven't seen it in production. I'll do that as well.

Comment 3

[Bob Clary [:bc] (inactive)]

https://github.com/mozilla/mozbase/commit/7b9acf714ce7aa6611ae8df4b0bb751570347a14 with debug change.