RE: slink_cd-1.07 --- revisited....
On Wed, 24 Feb 1999, Jim Westveer wrote:
>> OK, I see your immediate problem - ncpfs depends on ipx, which is not on
>> the first disc. Checking on this, it's missing from slink1.useful. Add it
>> there and all should work. The new pkg-order stuff will fail if
>> dependencies on disc 1 are not met and will give a warning if the others
>> fail.
>
>YUP, THAT WAS IT... thanks!
OK, good.
>> Hmmm. The md5sum errors here suggest you will be prone to the same errors
>> as Gunnar Isaksson - if the Packages files on your mirror are out of sync
>> then packages may not be copied over. It looks like the master Packages
>> files are broken. <sigh>
>
>No comprende. You mean that if the md5sum fails, the file is NOT copied
>to the output disk ???? No Bueno!!! or, perhaps I do not understand.
Not quite, no. But the failed md5sum is a symptom of a worse problem. The
Packages file in the mirror is about the only source of information about
the files and md5sums in the binary section. This means if it is not in
sync with the actual packages in the archive, 2 things will happen:
1 the md5sum stage will fail (it will be looking for different versions
for the packages) - this is annoying, but not really fatal
2 the genlist stage will also fail - genlist starts with a list of package
names only and uses the Packages file to get a filename for each one. If
this in not in sync with the mirror, then genlist will be listing a
different filename for some packages than those that you have. Result:
package is omitted. Aaaaargh!
There is a way around this, I guess. We can generate our own Packages file
from the files currently in the mirror. But this will only be any use for
(2) above - clearly it will be no use for md5sum checking, as we'll be
comparing against md5sums we've just created.
<sigh> It looks like I'll have to add another Packages step to slink_cd
now. This stuff is taking longer and longer... Maybe I'll be able to work
with timestamps - only generate the new Packages file if there are newer
files in the mirror. Hmmm, that could work.
>My mirror should be 'a good copy', as I just updated it from
>www.uk.debian.org this morning. (8:am PST) and the timestamps on...
>
>....oops it says "ls-lR 4886 Kb Mon Feb 22 22:46:00 1999"
>this be no good either...humm..I notice now that uk.debian has
>Feb 23 22:29:00 1999
>
>I will run the update overnight, and see what pops up tomorrow morning .
OK, great. I'm getting a bit worried that some of these are only showing
up so late...
--
Steve McIntyre, CURS CCE, Cambridge, UK. stevem@chiark.greenend.org.uk
Getting a SCSI chain working is perfectly simple if you remember that there
must be exactly three terminations: one on one end of the cable, one on the
far end, and the goat, terminated over the SCSI chain with a silver-handled
knife whilst burning *black* candles. --- Anthony DeBoer
Reply to: