Fedora Core 3 Test 3 available

John Mellor john at mellor.dyndns.org
Wed Oct 13 00:31:10 UTC 2004


On Tue, 2004-10-12 at 00:06, seth vidal wrote:
> On Mon, 2004-10-11 at 23:20 -0400, John Mellor wrote:
> > On Mon, 2004-10-11 at 10:25, Bill Nottingham wrote:
> > > Fedora Core 3 test 3 is now available.
> > 
> > Is the MD5sum correct as distributed?  I just did a completely
> > successful download using BitTorrent, and it says that everything was
> > ok.
> > 
> > According to the MD5 file, for the 5 disks, I should therefore see:
> > 
> >     bc4efae0c3db8f32820b5a779455d20f  FC3-test3-i386-disc1.iso
> >     2e925a39a949401abd57a26f0d09a672  FC3-test3-i386-disc2.iso
> >     09d7a1cfe6923b184987d5ecf5a82d8a  FC3-test3-i386-disc3.iso
> >     ec5fbf3381cdf718c7f50362740ce784  FC3-test3-i386-disc4.iso
> >     9ecbc904c654edec49666098082c3020  FC3-test3-i386-rescuecd.iso
> > 
> > But instead, I see:
> > 
> > !   4b1c867939bcc6bfde0c2396a276d3aa  FC3-test3-i386-disc1.iso
> >     2e925a39a949401abd57a26f0d09a672  FC3-test3-i386-disc2.iso
> > !   ef112940cfbaa9f64da52ec6e06db8ac  FC3-test3-i386-disc3.iso
> >     ec5fbf3381cdf718c7f50362740ce784  FC3-test3-i386-disc4.iso
> > !   c9e0f4dd7223b8819f1bc110b705d051  FC3-test3-i386-rescuecd.iso
> > 
> > which looks like bad disks 1, 3 and the rescue CD.  I cross-checked the
> > MD5SUM file contents with 2 other mirror sites and the master site, and
> > they all agree.    I just restarted my torrent, and it re-checked the
> > as-downloaded files and again found no fault.
> > 
> > RedHat has published bad MD5sums before, so that may be the real
> > problem.  Before I go and start installing somebody's spiffy new cracked
> > version, which one is really correct -- the successful download, or the
> > RedHat calculation?

> Statistically which do you think is more likely? That one download
> failed or that red hat published the wrong md5sums that have been
> verified by hundreds of mirror sites all over the world?

I agree fully with you, which leads me to a very disturbing conclusion
that there is a major fault in at least the latest version of
BitTorrent.  Having encountered some unknown error and ignored it,
BitTorrent is not correctly detecting that some kind of error occurred,
and is representing the incorrectly finished download as ok, and maybe
much worse, is presenting this bad download for upload by other parties.

That seems like such a fundamental trio of errors that I'm having some
difficulty understanding why this trainwreck of a programme is being
used.  I certainly will not be able trust it again.

Is there a workaround, maybe using an earlier version of BitTorrent? 




More information about the test mailing list