#10489 NORM F14: XO-1 (F14) - If not responded to, 'mission control' spawns and spawns

[Zarro Boogs per Child]

#10489: XO-1 (F14) - If not responded to, 'mission control' spawns and spawns
--------------------------+-------------------------------------------------
 Reporter:  mikus         |                 Owner:                                   
     Type:  defect        |                Status:  new                              
 Priority:  normal        |             Milestone:  F14                              
Component:  not assigned  |               Version:  Development build as of this date
 Keywords:                |           Next_action:  never set                        
 Verified:  0             |   Deployment_affected:                                   
Blockedby:                |              Blocking:                                   
--------------------------+-------------------------------------------------
 xo-1 build os3

 When I respond in a timely fashinon to the request for password that gets
 asked in a dialog when the Sugar session starts -- the xo-1 f14 builds
 start Sugar (allowing me to use the system).

 But this time I happened to push 'power on' - and walked away.  When I
 came back 20 minutes later, the XO-1 was repeating: (starting sugar,
 failing;  starting sugar, failing;  etc;  etc ...).

 HOWEVER -- everything in the XO-1 was being done __extremely slowly__.
 That gave me time to start up a text console.  'top' showed me perhaps a
 dozen instances of 'mission-control' (as it was dispatched, __each__ of
 those mission control tasks immediately conusmed ALL available CPU cycles
 !!) (that is, 'top' wait+idle == 0%).  Between instances of "loss of
 focus" (at each sugar restart) I tried to 'kill' those mission-control
 processes.  But they kept spawning almost as fast as I was killing them.
 I gave up and brute-forced a system dump.

 [[BR]]I don't know if that system dump will be very helpful - yes, it does
 show those 'mission-control' processes that I did not manage to kill.  But
 it seemed to me as though Sugar (and X-windows) were failing because they
 could not get access to their normal resources -- the reason they could
 not was because 'dbus' calls were not being responded to -- the reason for
 dbus not responding was because the 'mission-control' processes were
 eating up just about every possible CPU cycle.
 ----
 Note:  I've seen the "multiple mission-control processes (each trying to
 consume 100% of the CPU)" situation with xo-1 F14 build os2 as well.

-- 
Ticket URL: <http://dev.laptop.org/ticket/10489>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system