| Summary: | # urpmi --auto-update Could not find a mirror from mirrorlist $MIRRORLIST | ||
|---|---|---|---|
| Product: | Infrastructure | Reporter: | Palm Pre <palm_pre_stl> |
| Component: | Others | Assignee: | Sysadmin Team <sysadmin-bugs> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | major | ||
| Priority: | High | CC: | mageia, nanardon, rdalverny, sysadmin-bugs, thierry.vignaud |
| Version: | unspecified | Keywords: | Triaged |
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | urpmi | CVE: | |
| Status comment: | |||
| Attachments: |
Copy of /etc/urpmi.urpmi.cfg
Bug confirmed using clean Mageia 2 alpha 2 install |
||
|
Description
Palm Pre
2011-12-11 16:42:08 CET
was the network connected ? also please *attach* the file /etc/urpmi/urpmi.cfg CC:
sysadmin-bugs =>
(none) I can use Firefox to go online and post this respond Created attachment 1223 [details]
Copy of /etc/urpmi.urpmi.cfg
and urpmi --auto-update --debug ?
Manuel Hiebel
2011-12-11 21:12:36 CET
Attachment 1223 mime type:
application/octet-stream =>
text/plain # urpmi --auto-update --debug getting lock on urpmi parsing: /etc/urpmi/mediacfg.d/Devel-1-x86_64 parsing: /etc/urpmi/mediacfg.d/Devel-2-alpha1-x86_64 loading mirrors cache not using outdated cached mirror list $MIRRORLIST URPMI_ADDMEDIA_REASON reason=update getting mirror list from http://mirrors.mageia.org/api/mageia.2.x86_64.list?reason=update retrieving mageia.2.x86_64.list?reason=update '/usr/bin/wget' '--force-clobber' '-q' '--retr-symlinks' '--timeout=60' '-P' '/var/cache/urpmi/partial' 'http://mirrors.mageia.org/api/mageia.2.x86_64.list?reason=update' | retrieved mageia.2.x86_64.list?reason=update found geolocalisation US 41.85 -86.35 from timezone America/Chicago Could not find a mirror from mirrorlist $MIRRORLIST unlocking urpmi database EXITING (pid=6445) (In reply to comment #5) > # urpmi --auto-update --debug > getting lock on urpmi > parsing: /etc/urpmi/mediacfg.d/Devel-1-x86_64 > parsing: /etc/urpmi/mediacfg.d/Devel-2-alpha1-x86_64 > loading mirrors cache > not using outdated cached mirror list $MIRRORLIST > URPMI_ADDMEDIA_REASON reason=update > getting mirror list from > http://mirrors.mageia.org/api/mageia.2.x86_64.list?reason=update > retrieving mageia.2.x86_64.list?reason=update > '/usr/bin/wget' '--force-clobber' '-q' '--retr-symlinks' '--timeout=60' '-P' > '/var/cache/urpmi/partial' > 'http://mirrors.mageia.org/api/mageia.2.x86_64.list?reason=update' | > retrieved mageia.2.x86_64.list?reason=update > found geolocalisation US 41.85 -86.35 from timezone America/Chicago > Could not find a mirror from mirrorlist $MIRRORLIST > unlocking urpmi database > EXITING (pid=6445) Thierry any ideas ? @Palm Pre, its a new install or upgrade from mageia 1 ? Keywords:
(none) =>
Triaged
Manuel Hiebel
2011-12-11 22:46:47 CET
Severity:
normal =>
major A new install. Created attachment 1224 [details]
Bug confirmed using clean Mageia 2 alpha 2 install
http://mirrors.mageia.org/api/mageia.2.x86_64.list?reason=upgrade and the like now return nothing CC:
(none) =>
sysadmin-bugs
Thierry Vignaud
2011-12-12 08:55:36 CET
CC:
(none) =>
thierry.vignaud ah yes indeed it's work with s/2/cauldron Sysadmins, please fix before the alpha2 Priority:
Normal =>
High FYI, it's due to fixing the version in /etc/product.id that violates prior usage. Back in february, version was altered in mageia-release from 1/2/3/... to Cauldron whereas we never set it to cooker back @mdv (was hidden under "only obsolete mandriva-release-* to allow clean upgrade" commit message). This was bogus as this field is used by quite some tools and broke them (was partially hidden due to wiki not being ready). We should stick to historical behavior (last release pointing to cauldron, prior/released releases pointing to proper stable tree). as a workaround use: urpmi.addmedia --distrib --mirrorlist "http://mirrors.mageia.org/api/mageia.cauldron.$ARCH.list' replacing $ARCH by i586 or x86_64. # urpmi.addmedia --distrib --mirrorlist 'http://mirrors.mageia.org/api/mageia.cauldron.$x86_64.list' adds medias however rsync://distro.ibiblio.org/mageia/distrib/cauldron/x86_64/media/core/release/media_info/20111212-093152-synthesis.hdlist.cz ...retrieving failed: rsync failed: exited with 23 problem reading synthesis file of medium "Core Release" (In reply to comment #11) > This was bogus as this field is used by quite some tools and broke them (was > partially hidden due to wiki not being ready). Do we have documentation about it now? (In reply to comment #14) > So what is needed is to add Mageia 2 to https://mirrors.mageia.org ? We need at least http://mirrors.mageia.org/api/mageia.2.i586.list (and x86_64) listing available mirrors (do we have something checking mirrors sync btw or not really?). CC:
(none) =>
rdalverny The way we did it back at Mdv was: when we release a new distro, we: - switched api.<vendor>.com/20XX.X to stable tree - added new 20XX+1.X to api.<vendor>.com - bumped the version in <vendor>-release.rpm and uploaded it in cooker - open the cooker tree again Not so many people remembers this but blino is one of them CC:
(none) =>
mageia Ok, so we can do the same. mirrors list for mageia 2 is now returning the list of cauldron mirrors : http://mirrors.mageia.org/api/mageia.2.x86_64.list (In reply to comment #15) > listing available mirrors (do we have something checking mirrors sync btw or > not really?). At the moment no. See bug #2844. In current version of MGA::Mirrors, the only check done is to try to fetch the URLs for all mirrors, protocols, distributions, architectures, as seen in check_distributions function in : http://svnweb.mageia.org/soft/mirrors/mga-mirrors/trunk/lib/MGA/Mirrors/DB.pm?revision=1619&view=markup It would be nice to also check the date inside timestamp file. And even better would be to also check the signature of the hdlists (but first, we have to sign them). (In reply to comment #17) > Ok, so we can do the same. > > mirrors list for mageia 2 is now returning the list of cauldron mirrors : > http://mirrors.mageia.org/api/mageia.2.x86_64.list Thanks ! (so we can close ?) Yes, I think we can close. Status:
NEW =>
RESOLVED
Nicolas Vigier
2014-05-08 18:07:04 CEST
CC:
boklm =>
(none) |