Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 923229 - rsync mirror rsync12.de.gentoo.org is out of date
Summary: rsync mirror rsync12.de.gentoo.org is out of date
Status: RESOLVED FIXED
Alias: None
Product: Mirrors
Classification: Unclassified
Component: Server Problem (show other bugs)
Hardware: All Linux
: Normal normal
Assignee: Mirror Admins
URL: rsync://rsync12.de.gentoo.org/gentoo-...
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-01-29 08:09 UTC by SigHunter
Modified: 2024-02-12 12:10 UTC (History)
5 users (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 SigHunter 2024-01-29 08:09:04 UTC
The rsync mirror rsync://rsync12.de.gentoo.org/gentoo-portage/ is out of date and syncing with it and emerge updating downgrades installed packages.

Manifest timestamp: 2024-01-26 08:10:16 UTC

I tried to contact the listed admin Timo Rothweiler <tr.bgo@gmx.de> but only got a smtp error 'mailbox unavailable'

Reproducible: Always

Steps to Reproduce:
1. set sync-uri = rsync://rsync.de.gentoo.org/gentoo-portage or rsync.europe.gentoo.org in repos.conf/gentoo.conf
2. emaint --sync
3. emerge -uND @world 
Actual Results:  
Old repo data will downgrade newer installed packages

Expected Results:  
all mirrors in sync
Comment 1 SigHunter 2024-01-29 08:17:50 UTC
Correction, reproducible steps must of course be: sync-uri = rsync://rsync12.de.gentoo.org/gentoo-portage/
Comment 2 Christian Zuckschwerdt 2024-02-09 11:51:08 UTC
I can confirm that the rsync mirror is out of date for me.

Is one of the mirror admins already looking into this? Or should we try and contact someone at amiconsult GmbH, Karlsruhe (sponsor of the mirror)?
Comment 3 Bernd Feige 2024-02-10 08:53:54 UTC
Confirm - also already since 2024-01-29, each time I am unlucky enough to hit this mirror, my system starts downgrading to something several weeks ago... I wonder why there is no safeguard against this - just as the signature is checked, the timestamp could be checked as well to disallow rsync with an earlier tree than the current by default...
Comment 4 Bernd Feige 2024-02-10 08:59:07 UTC
Tried to send an E-Mail to the address given in the rsync message:
Timo Rothweiler <tr.bgo@gmx.de>

But that mailbox is dead:

tr.bgo@gmx.de:
        SMTP error from remote server for RCPT TO command, host: mx01.emig.gmx.net (212.227.17.5) reason: 550 Requested action not taken: mailbox unavailable
Comment 5 SigHunter 2024-02-10 09:06:47 UTC
portage even checks for the age of the manifest and throws a warning
based on sync-rsync-verify-max-age = x days in repos.conf

>>> Starting rsync with rsync://[2a01:4f8:140:5123::2]/gentoo-portage...
 * Manifest is over 3 days old, this is suspicious!
 * You may want to try using another mirror and/or reporting this one:
 *   rsync://[2a01:4f8:140:5123::2]/gentoo-portage
 * 
 * Manifest timestamp: 2024-01-26 08:10:16 UTC

that warning should be (configurable as) an error and skip the current rsync mirror and retry with another, IF that is possible with rsync at this stage of sync
Comment 6 Robin Johnson archtester Gentoo Infrastructure gentoo-dev Security 2024-02-10 23:04:16 UTC
DNS entries for rsync12.de removed
Comment 7 Sam James archtester Gentoo Infrastructure gentoo-dev Security 2024-02-12 12:10:58 UTC
(In reply to SigHunter from comment #5)
> portage even checks for the age of the manifest and throws a warning
> based on sync-rsync-verify-max-age = x days in repos.conf
> 
> >>> Starting rsync with rsync://[2a01:4f8:140:5123::2]/gentoo-portage...
>  * Manifest is over 3 days old, this is suspicious!
>  * You may want to try using another mirror and/or reporting this one:
>  *   rsync://[2a01:4f8:140:5123::2]/gentoo-portage
>  * 
>  * Manifest timestamp: 2024-01-26 08:10:16 UTC
> 
> that warning should be (configurable as) an error and skip the current rsync
> mirror and retry with another, IF that is possible with rsync at this stage
> of sync

I think it is a nerror as it is, as it fails verification? Maybe the output isn't ideal though, please file a new bug for that if you wouldn't mind.