forcemerge 624122 616064 616489
tag 624122 unreproducible moreinfo
thanks
On Mon, Apr 25, 2011 at 21:06, <jidanni@jidanni.org> wrote:
> It happens about 30% of the time when I do apt-get update.
Never happened to me so far, it seems i have an extraordinary amount of luck. :)
The issue itself is not really a problem of APT but of the mirror
(or your proxy, or or or), but lets ignore that for a minute:
So, i wrote a testscript puzzled together from the information you
provided in the three bugreports:
The script is attached and can be placed in the source tree into
test/integration for execution, but here is a commented run:
First some boilerplate: I am creating a 'somepkg' v1 package in a
fresh repository, clear var/lib/apt/lists just to be sure and run
$ apt-get update
Get:1 http://localhost unstable InRelease [3583 B]
Get:2 http://localhost unstable/main Sources [438 B]
Get:3 http://localhost unstable/main i386 Packages [375 B]
Get:4 http://localhost unstable/main TranslationIndex [1504 B]
Get:5 http://localhost unstable/main Translation-en [279 B]
Fetched 6179 B in 0s (36.0 kB/s)
As expected everything is fine.
Now i am creating 'somepkg' v2 for the archive, update all
indexes but the InRelease file and run again:
$ apt-get update
Reading package lists...
Get:1 http://localhost unstable InRelease [3583 B]
Get:2 http://localhost unstable/main Sources [672 B]
Get:3 http://localhost unstable/main i386 Packages [551 B]
Get:4 http://localhost unstable/main TranslationIndex [1504 B]
Get:5 http://localhost unstable/main Translation-en [349 B]
Fetched 6659 B in 0s (39.0 kB/s)
W: Failed to fetch
bzip2:/tmp/tmp.NO42yF6oSu/rootdir/var/lib/apt/lists/partial/localhost:8080_dists_unstable_main_source_Sources
Hash Sum mismatch
W: Failed to fetch
bzip2:/tmp/tmp.NO42yF6oSu/rootdir/var/lib/apt/lists/partial/localhost:8080_dists_unstable_main_binary-i386_Packages
Hash Sum mismatch
E: Some index files failed to download. They have been ignored, or old
ones used instead.
Oh no, an error! But expected so far, as the InRelease file is
outdated as it should to reproduce your bugreport.
Now lets look into the var/lib/apt/lists directory:
$ find var/lib/apt/lists
rootdir/var/lib/apt/lists
rootdir/var/lib/apt/lists/localhost:8080_dists_unstable_main_i18n_Translation-en
rootdir/var/lib/apt/lists/localhost:8080_dists_unstable_main_i18n_Index
rootdir/var/lib/apt/lists/localhost:8080_dists_unstable_InRelease
rootdir/var/lib/apt/lists/localhost:8080_dists_unstable_main_binary-i386_Packages
rootdir/var/lib/apt/lists/localhost:8080_dists_unstable_main_source_Sources
rootdir/var/lib/apt/lists/partial
rootdir/var/lib/apt/lists/partial/localhost:8080_dists_unstable_main_binary-i386_Packages.decomp.FAILED
rootdir/var/lib/apt/lists/partial/localhost:8080_dists_unstable_main_source_Sources.decomp.FAILED
rootdir/var/lib/apt/lists/partial/localhost:8080_dists_unstable_main_binary-i386_Packages
rootdir/var/lib/apt/lists/partial/localhost:8080_dists_unstable_main_source_Sources
Huh? Still the old files around in lists/ ?
And especially no FAILED file outside of partial/
Lets check that:
$ apt-cache policy somepkg
somepkg:
Installed: (none)
Candidate: 1.0
Version table:
1.0 0
500 http://localhost/ unstable/main i386 Packages
$ apt-cache show somepkg
Package: somepkg
Priority: optional
Section: others
Installed-Size: 16
Maintainer: Joe Sixpack <joe@example.org>
Architecture: i386
Version: 1.0
Filename: pool/somepkg_1.0_i386.deb
Size: 1230
MD5sum: 4e96f7fab53075a85d791fbdedc2d683
SHA1: cbdf42cd9a601132333670828143a9400b145ecc
SHA256: e8bf9c163673cb9c7ad0079245830eadcf6e206100a78ba996c1bfde98decf89
Description-en: an autogenerated dummy somepkg=1.0/unstable
If you find such a package installed on your system,
YOU did something horribly wrong! They are autogenerated
und used only by testcases for APT and surf no other propose…
Description-md5: 8078bca1645dcc8c5a94f3a1116c6005
Looks like everything is as it should be…
So, could you tell us a bit more about your setup, used mirror and
especially what you did so far to "fix" it (most of the time, thats
the problem).
I currently don't see a way in the source for derivation as this message
(in this context) is printed only at a single location in the code…
Best regards
David Kalnischkies
Attachment:
test-bug-624122-hash-sum-mismatch
Description: Binary data