#10019 NORM Not Tri: datastore service killed by presence of external SD card at boot

[Zarro Boogs per Child]

#10019: datastore service killed by presence of external SD card at boot
------------------------------------+---------------------------------------
           Reporter:  gypsymoth     |       Owner:               
               Type:  defect        |      Status:  new          
           Priority:  normal        |   Milestone:  Not Triaged  
          Component:  not assigned  |     Version:  not specified
         Resolution:                |    Keywords:               
        Next_action:  never set     |    Verified:  0            
Deployment_affected:                |   Blockedby:               
           Blocking:                |  
------------------------------------+---------------------------------------

Comment(by Quozl):

 Corrections.

 The Terminal log shows a failure to connect to the datastore via D-Bus.
 Was it a journal resume (coloured icon) or a new activity (monochrome
 icon)?

 I've installed an 802, checked, and os.stat in backingstore.py is already
 protected by an exception handler, which places the exception description
 in the datastore.log but in theory does not terminate the datastore
 process.

 Therefore the datastore process is either present and not responding (busy
 reading the files on your 16GB card all the way through first), or has
 stopped running, or there is a failure of D-Bus.

 The sugar.log shows a failure of D-Bus for communication with
 NetworkManager, and later an introspection timeout.  It is as if D-Bus is
 broken as well.  But D-Bus processes are present.  Perhaps the timeout
 happened because of CPU resource stolen by datastore.

 ps.out in your logs also shows pid 1293, the datastore-service, is in
 state D (waiting on I/O to complete), so chances are the reason why
 Terminal could not be launched was that the datastore service was locked
 up reading your no-more-than 16Gb of content.

 It would be interesting to find out if the same 16Gb SD card affects the
 more recent XO-1.5 in the same way.

-- 
Ticket URL: <http://dev.laptop.org/ticket/10019#comment:12>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system