This post would attempt to wrap up any lingering doubts on using Jigdo Jigsaw. Some issues I saw and how they were corrected. Also something about a future release/roadmap if some developer worked on it.
while the last time we were freshening the archive, this time we shall see what kind of issue you can get. Sometimes you can get this kind of error :-
FINISHED --2010-12-27 16:39:31--
Downloaded: 10 files, 2.4M in 1m 16s (31.5 KB/s)
Found 10 of the 39 files required by the template
Error: `debian-testing-amd64-DVD-1.iso.tmpdir/www.mirrorservice.org/sites/ftp.debian.org/debian/pool/main/j/java-common/default-jre_1.6-40_amd64.deb' does not match checksum in template data
Error: `debian-testing-amd64-DVD-1.iso.tmpdir/www.mirrorservice.org/sites/ftp.debian.org/debian/pool/main/b/bogofilter/bogofilter_1.2.2-2_amd64.deb' does not match checksum in template data
Error: `debian-testing-amd64-DVD-1.iso.tmpdir/www.mirrorservice.org/sites/ftp.debian.org/debian/pool/main/a/auto-install/auto-install-nonet_1.9_all.udeb' does not match checksum in template data
Error: `debian-testing-amd64-DVD-1.iso.tmpdir/www.mirrorservice.org/sites/ftp.debian.org/debian/pool/main/g/gcc-defaults/gcj-jre_4.4.5-1_amd64.deb' does not match checksum in template data
Copied input files to temporary file `debian-testing-amd64-DVD-1.iso.tmp' - repeat command and supply more files to continue
Now I do not know why this happens/happened but probably due to mismatch between the copies on the various servers.
From what little I know of the Debian Infrastructure, it seems the main server is in Europe somewhere (don’t remember offhand whether Germany, Poland or some other country) and it takes time for the rest of the mirrors to get the newer/freshest archive. So sometimes if you have the latest database (the latest .template file) but the mirror from which you are trying to download has still the earlier files only then you will see something like above.
The only workaround I was able to manage/do is to go to the us or the original debian.org and download the files from there. Its easier for self and for the main mirror if I’m hitting it for 100 MB or thereabouts (depends on the changes in the mirror) then to hit the main one every week for a whole 4.4 GB.
Note: I have seen most of the servers in my country I have to go about the archive in a round-about way. A simple traceroute shows that most of the times I have to Singapore or whatever and then come back. Also most of the servers are not up all the time, nor are they current so cannot use them.
One of the more interesting things that should happen in the near future the files will be much more compressed using xz utils than it is today meaning saving bandwidth all around with similar user experience as today on the user’s desktop’s as far as upgrades go. Would be looking forward to that.
Here is my wish list if jigdo-jigsaw would be developed again :-
a. Have support for SHA1 or better SHA256 :- There have been known of shortcomings of md5sum and attacks on md5sum some years ago. Searching on Google or on Wikipedia should give more info. on that. In that reference SHA256 would be more robust.
b. Have some kind of support to apt-p2p or something similar :- The problem I see is that jigdo is good only when the server is good and is dependant on the server infrastructure. The best would be to have some kind of p2p interface so if I download from the server some packages, then the other guy can download from me and/or vice-versa.
While the first one should be easy, the second one would be a bit more complex.All in all a very useful and interesting tool, bit sad that it’s in maintenance though 😦
That’s all for now, see you all next week 🙂