canonical-ci-engineering team mailing list archive

[Vincent Ladeuil <vila+ci@xxxxxxxxxxxxx>]

>>>>> Andy Doan <andy.doan@xxxxxxxxxxxxx> writes:

    > On 11/05/2013 03:19 AM, Vincent Ladeuil wrote:
    >> I can think of two paths from here:
    >> 
    >> - revisit the decision to flash devices so often and investigate how to
    >> ensure we can provide a clean environment (can we get help from
    >> system-image to setup an arbitrary image from inside the phone instead
    >> of flashing a whole image?),

See the other thread started by Francis about Saviq's suggestion, I
couldn't find back where he mentioned that in my mails (so may be he
only did it on IRC). I'll comment further in that other thread.

    >> 
    >> - stress testing*locally*  to get a better understanding on when and why
    >> the devices get stucked. Evan gave a link for an alternate connection
    >> through the audio jack but it's for the mako. Is there such a thing
    >> for the maguro ?

    > I'm not sure that option solves *this* problem.

Right, it's not about solving the problem, it's about diagnosing it.

    > It might, but I'm not sure having a serial console to a device
    > that is stuck/oops'd/etc will help.

I have no idea about what can comes out of that, that's the point, I'm
searching for more data.

          Vincent