Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 92041 - Some source tarballs failling md5 checks on http://ftp.uni-erlangen.de/pub/mirrors/gentoo/
Summary: Some source tarballs failling md5 checks on http://ftp.uni-erlangen.de/pub/mi...
Status: RESOLVED FIXED
Alias: None
Product: Gentoo Infrastructure
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All All
: High major (vote)
Assignee: ftp.uni-erlangen.de mirror role account
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-05-09 13:22 UTC by Benjamin Smee (strerror) (RETIRED)
Modified: 2005-05-13 08:24 UTC (History)
1 user (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Benjamin Smee (strerror) (RETIRED) gentoo-dev 2005-05-09 13:22:37 UTC
I have had a few failed md5 checks on various source tarballs tonight from the http://ftp-stud.fht-esslingen.de/pub/Mirrors/gentoo/ mirror. Apparently some others has reported issues in #gentoo-security. I think it should be investigated further, I will update this bug with more packages that fail md5 checks when I have a chance to check some logs.
Comment 1 Benjamin Smee (strerror) (RETIRED) gentoo-dev 2005-05-09 13:24:40 UTC
sed is also failling md5 checks
Comment 2 Benjamin Smee (strerror) (RETIRED) gentoo-dev 2005-05-09 13:32:20 UTC
!!! Digest verification Failed:
!!!    /usr/portage/distfiles/Linux-PAM-0.77.tar.gz
!!! Reason: Failed on MD5 verification

Comment 3 Lance Albertson (RETIRED) gentoo-dev 2005-05-09 13:38:37 UTC
I'm wondering if this is a bad sync from the mirror.

FYI, found this info for the admin:

This is an experimental FTP server. If you have any unusual problems,
please report them via e-mail to adrian@fht-esslingen.de and we will 
do our best to help you.

That email isn't in bugzilla or I'd cc: it!
Comment 4 Benjamin Smee (strerror) (RETIRED) gentoo-dev 2005-05-09 14:05:45 UTC
http://ftp.uni-erlangen.de/pub/mirrors/gentoo/ is the mirror I meant to paste. I had the other one in buffer when I posted originally, sorry about the confusion
Comment 5 Lance Albertson (RETIRED) gentoo-dev 2005-05-09 15:10:46 UTC
Thanks, found the admin email.. adding...
Comment 6 ftp.uni-erlangen.de mirror role account 2005-05-09 22:55:43 UTC
We had some outtime (as mentioned at gentoo-mirrors list), probably you have downloaded while the initial sync was in progress.
I got my initial sync (cause of high speed) from mirrors.sec.informatik.tu-darmstadt.de but i just resynced to the master.
Sorry for any inconveniences. If this problem consists please report again.

Comment 7 ftp.uni-erlangen.de mirror role account 2005-05-10 11:55:50 UTC
We seem to have some serious hardware problem. A detailed check of _all_ files in /distfiles by verifying them with md5sum against the digests recorded in portage gave many broken files.

Its currently unknown where these come from (disk, net, cpu, ram, ...). We try to find out as soon as possible.

For now, I unmounted the _complete_ gentoo mirror to prevent further problems for the users.

I will update this bug-report frequently.

mirror-admin: please assign this bug to the ftp.uni-erlangen.de role account - I will take care.
Comment 8 Lance Albertson (RETIRED) gentoo-dev 2005-05-10 12:05:49 UTC
Ok, thanks for the update. Going to remove infra since this clearly a mirror-only issue. 

Cheers
Comment 9 ftp.uni-erlangen.de mirror role account 2005-05-11 11:23:14 UTC
We tracked down this problem to:
https://www.redhat.com/archives/linux-lvm/2005-January/msg00067.html
(Keep something like this in mind if you have probs like this)

So we kicked LVM and just used the plan /dev/sda provided by the raid array.

We now run verify runs - but the first tests are positive.
Comment 10 ftp.uni-erlangen.de mirror role account 2005-05-13 08:24:53 UTC
After long testing no more errors occured.

All distfiles have been refetched and successfully verified against the md5sums listed in portage (even if this was a little "scripting-torture"... ;)

This issue should be resolved - Sorry for any inconveniences.