#9927 LOW Future : filesystem has trouble with large yum installs

[John Watlington]

On Dec 28, 2009, at 11:15 PM, Chris Ball wrote:

> Hi,
>
>> While #9927 is personally interesting to me, can I instead divert
>> your attention to the logs I just posted for #9865 ?
>
> Looks identical to the bug (#9702) with 3.3V not coming up fast enough
> to successfully resume the internal SD card.  We should check that
> your C1 brings up the 3.3V rail quickly enough.
>
>> Tmw. I will be in a position to verify this bug on a largish
>> number of C1 machines, and can hand off the poster child
>> (prebuild #2) to someone else.
>>
>> I would blame hardware, but why does it only happen when OHMD is
>> doing the suspending ?
>
> If we convince ourselves that it really does only happen with ohm, and
> also only with rtcwake mixed with autowack, and doesn't happen on B2s
> in the same configuration, we might find that we have #1835-like
> problems when mixing autowack with an ohmd rtcwake immediately after
> suspend hits -- we can get a resume happening below the EC's
> autowack-delay in that case, which is known to cause problems.
>
> I suppose the answer in that case would be "don't do that".
> Richard, what do you think?

Naw.  I say we hunt it down and kill it.

Probably a rail that needs a little help decaying.

wad