#9964 LOW Not Tri: 1st boot shallow copy needs feedback

Zarro Boogs per Child bugtracker at laptop.org
Tue Jan 12 01:00:43 EST 2010


#9964: 1st boot shallow copy needs feedback
-----------------------------------+----------------------------------------
           Reporter:  rsmith       |       Owner:  cjb          
               Type:  enhancement  |      Status:  new          
           Priority:  low          |   Milestone:  Not Triaged  
          Component:  distro       |     Version:  not specified
         Resolution:               |    Keywords:               
        Next_action:  never set    |    Verified:  0            
Deployment_affected:               |   Blockedby:               
           Blocking:               |  
-----------------------------------+----------------------------------------

Comment(by Quozl):

 I've tested the initial boot of current builds for 10.2, and the shallow
 copy is certainly not present.

 I went looking for other delays.

 There was an eight second delay in preparing SSH host keys.  We could
 avoid that by not including openssh-server in the build, or by configuring
 it off (chkconfig sshd), or by setting AUTOCREATE_SERVER_KEYS to NO in
 /etc/sysconfig/sshd.

 I don't know of anything else in our stack (e.g. Sugar, olpc-update) that
 depends on sshd, but I'd dislike losing it altogether because I use it for
 access to test units.

 Here's some results, a first boot followed by a forced power down at the
 nick prompt, a second boot to the nick prompt, a third and fourth boot to
 the activity ring.

 ||''task''||''ssh keygen begin''||''ssh keygen end''||''sugar waits for
 input''||
 ||first boot||39.46 sec||51.45 sec||70.56 sec||
 ||second boot||38.30 sec||50.96 sec||60.59 sec||
 ||third boot||-||-||38.96 sec||
 ||fourth boot||-||-||40.11 sec||
 ||restarting sugar||-||-||9.33 sec||

 The second boot also included an SSH keygen, probably because the power
 down happened before the filesystem was updated from the first keygen.

-- 
Ticket URL: <http://dev.laptop.org/ticket/9964#comment:2>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system


More information about the Bugs mailing list