[Sugar-devel] RFH - Journal corruption reports fom 8.2.1

[Martin Langhoff]

On Fri, Aug 21, 2009 at 12:13 AM, Jim Simmons<nicestep at gmail.com> wrote:
> For what it's worth, I've experienced this running the Sugar test
> environment in both Fedora 10 and 11.  It would not seem to be
> specific to .82.  The way to reproduce the problem is to run the Sugar
> environment in Xephyr (sugar-emulator).  Do some stuff, then instead
> of shutting down cleanly click on the "close" button on the Xephyr
> window.  That will close your Sugar session and practically guarantee
> that the next time you open it your Sugar Journal will be empty.  I
> regret to say that I've done this a few times.

That's not very encouraging. I "switch off" my XO, running 8.2.x
(which is 0.82.x of Sugar) all the time by removing the battery, which
is a much harsher method than what you describe (which is merely a
SIGTERM).

Sugar 0.84 has seen a reimplementation of the Datastore (by Tomeu?)
and I am sure he'll be interested in this. Avoiding dataloss and
generally being atomic is always one of the most important aspects of
the DS.

I can't cast the first stone[1] but I do want to say: any
Journal-related dataloss is a major concern. Let's report it, and
diagnose it. I have only seen dataloss on SoaS, but related to vfat /
overlay corruption (widely discussed).

cheers,




m
1 - The very early XS images _repartition & format your hard drive
soon after booting, without asking for permission_. I fixed it as soon
as I could move to a better installer. Still, it managed to eat
someone's HD. Shame on me.
-- 
 martin.langhoff at gmail.com
 martin at laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff