#8711 BLOC Not Tri: Sugar-shell burning up cumulative 1-2% of CPU resources after every Activity launch/resume
Zarro Boogs per Child
bugtracker at laptop.org
Sun Sep 28 16:24:39 EDT 2008
#8711: Sugar-shell burning up cumulative 1-2% of CPU resources after every
Activity launch/resume
-------------------------+--------------------------------------------------
Reporter: garycmartin | Owner: marco
Type: defect | Status: new
Priority: blocker | Milestone: Not Triaged
Component: sugar | Version: Development build as of this date
Keywords: blocks:8.2? | Next_action: never set
Verified: 0 | Blockedby:
Blocking: |
-------------------------+--------------------------------------------------
Testing on build 8.2-765, and joyride-2491.
Sugar-shell is burning up a cumulative of 1-2% of CPU resources after each
Activity launch or Journal resume.
Example: Launch and then quit Terminal 10 times from the Home view. The
sugar-shell process is usually at 0.1% CPU when left to idle after a
reboot - after 10 Activity launches (quitting each after it's started) and
allowing the system to idle, 18-20% of your CPU cycles will still be being
eaten by sugar-shell. The effect is cumulative, after running/resuming and
quitting another 10 Terminal activities, sugar-shell will be sitting at
idle consuming 43% of CPU resources (at which point things are really
miserable).
So far I've tested Terminal, Log, TurtleArt, Moon, Chat, Write and
observed this issue when working through a list of activities to smoke
test – noticed I could get through about 6-10 Activities (one launch and
one resume each) before needing to restart sugar.
Just to be clear, no activities need to be left running, so this is
perhaps the kind of work flow a kid might get into taking their XO from
class to class launching Activities, and stopping each after use, during a
school day – woe betide the kid who doesn't close activities...
Restarting Sugar with Ctrl-alt-delete gets the CPU resources back to the
99.x% idle.
Possibly what thomaswamm was seeing back in #8098, but no signal in that
ticket that it was related to cumulative launch/resumes that turns sugar-
shell into a CPU hog monster so he might have something different there
still.
--
Ticket URL: <http://dev.laptop.org/ticket/8711>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system
More information about the Bugs
mailing list