Rockbox.org home
release
dev builds
extras
themes manual
wiki
device status forums
mailing lists
IRC bugs
patches
dev guide



Rockbox mail archive

Subject: Re: Fwd:Jukebox Recorder stops working after 36 hours or so
From: Jean Boullier (jean.boullier_at_wanadoo.fr)
Date: 2003-07-17


Chris,
I agree that the probability that a JBR on its AC can run "for ever" is
high (it even has a power backup capability of several hours). But if my
professional credibility or my company's visibility were at stake I
would have tested at least one or two weeks to make sure that there is
no repeatable problem (e.g. after a certain number of repeat cycles),
that all files are clean, etc... If it is OK it is a nice cheap solution
to their need but frankly, one would never think a JBR is a professional
equipment and would never dare to blame Archos for the JBR not being
one. Also, although that would not solve the problem of week ends if the
box stops after 36 hours, I would add one line in the job duties of the
person in charge of the telephone system for doing an OFF-ON operation
once a day: that takes 30 seconds and is enough to eliminate possible
long run problems.

 From what I have learned of Linus since I read this list I am sure he
has taken a number of precautions in his 24/7 applications to protect
them from inopportune problems, e.g. serious error detection, error
recovery routines, periodical memory reset and probably a few other
tricks of his own, not forgetting serious testing and monitoring to
anticipate as much as possible.

I have personally been in charge of such applications and although a
short breakdown once in a while would not have been a real disaster I
was checking as many things as I could think of, especially those which
"will never happen" and I had an automatic restart once a day (in the
computers I used it was taking a few seconds) . All sorts of things that
one could not reasonably blame the poor Archos programmer for having not
put them in its firmware! He has already enough responsibility for real
mistakes : I bought my JBR6 in August 2001 and at this time the main
problem was not that the firmware was not clever, it was that a number
of its functions were not working at all!

Cheers. Jean.

-- 
Jean Boullier - Rouen - Normandie - France

Chris Holt wrote:

>On Thu, 17 Jul 2003 01:39:26 +0200, Jean Boullier wrote: > >>If the hospitals in your area buy their 24/7 equipments with the same >>level of "professionalism" as this guy I sincerely advise you to never >>need to go there.... 8-) >> >>Anyway with the poor problem report that this guy has given one cannot >>say that the JB stopped working, it probably just stopped and just >>needed to be restarted, as any DP system which has not been >>specidically designed for 24/7 usage. The MTBF of any HD nowadays is by >>far superior to the MTBF of a complete application including hardware >>and software if it has not been designed for an (apparently) continuous >>availability.. >> >>Personally I would not even give a single euro to a lawyer for filing >>such a case... >>Jean. >> > >I don't lend any weight to the case either. Warranties usually >specifically disclaim any liability for suitability or fitness for any >particular use. > >But I also feel compelled to discount your analogies. The Archos is >indeed more similar to bio-medical equipment than to DP systems owing to >it's specificity of purpose and relatively simplistic OS/firmware. > >As to DP systems, I believe that Linux has proven that many "complete >applications" used in the complex way of a general purpose computer can >indeed run 24/7 for years, even with off the shelf hardware and a >relatively complex OS and program suite, even though not specifically >designed to do so. > >I do believe the user should have simply rebooted the Jukebox and seen if >the problem repeated before contacting technical support. Could have just >been a corrupt MP3 for all anyone knows. But I see little reason other >than a lack of robustness in the firmware that the Archos could not be >counted on to run reliably on a 24/7 basis. This is why I encouraged the >user to give Rockbox a shot. Better buffering, more flexible charging, >drive spindown etc. combined with identification and elimination of >corrupted files should provide a very reliable system, especially when >considering that this is a stationary and presumably climate controlled >application. > >Chris >"And uh, get the machine that goes *PING*" > > >



Page was last modified "Jan 10 2012" The Rockbox Crew
aaa