Summary: | masked by corruption (for no apparent reason) | ||
---|---|---|---|
Product: | Portage Development | Reporter: | Mike Auty (RETIRED) <ikelos> |
Component: | Core - Interface (emerge) | Assignee: | Portage team <dev-portage> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | david.morgan, dm.konrad, jakub, wsim |
Priority: | High | Keywords: | InVCS, REGRESSION |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 147007 | ||
Attachments: |
'Corrupted' /var/cache/edb file
patch for debugging purposes Small debug log from failed mpc installation attempt |
Description
Mike Auty (RETIRED)
2006-12-20 03:57:19 UTC
D'oh, sorry, I forgot to include the exact symptoms: ikelos ~ # emerge -pv =mpc-0.12.0 These are the packages that would be merged, in order: Calculating dependencies !!! All ebuilds that could satisfy "=mpc-0.12.0" have been masked. !!! One of the following masked packages is required to complete your request: - media-sound/mpc-0.12.0 (masked by: corruption) For more information, see MASKED PACKAGES section in the emerge man page or refer to the Gentoo Handbook. Can't reproduce...
$ emerge -pv =mpc-0.12.0
These are the packages that would be merged, in order:
Calculating dependencies... done!
[ebuild N ] media-sound/mpc-0.12.0 USE="bash-completion nls" 0 kB
* checking ebuild checksums ;-) ... [ ok ]
* checking auxfile checksums ;-) ... [ ok ]
* checking miscfile checksums ;-) ... [ ok ]
* checking mpc-0.12.0.tar.bz2 ;-) ... [ ok ]
>>> Unpacking source...
>>> Unpacking mpc-0.12.0.tar.bz2 to /var/tmp/portage/media-sound/mpc-0.12.0/work
>>> Source unpacked.
http://forums.gentoo.org/viewtopic-t-525908.html Some more people with the same problem. Hmm, the forums seem to suggest it's a corruption of files in the /var/cache/edb directory (which I think is the VDB?). Is it worth attaching the contents of the files in there? Also, is there a way to regenerate these without diving in and deleting files directly? Seemingly a sync, and a metadata sync didn't recreate the data based on the ebuilds themselves, so I'm not sure how else to ask for the VDB to be ignored or recreated using portage. If anyone has anything they'd like me test I'll leave one machine exhibiting the problem for a while... *** Bug 158575 has been marked as a duplicate of this bug. *** I think this may be related to disk/filesystem trouble. If you're experiencing this, please try `rm -rf /var/cache/edb/dep && emerge --metadata` to see if that helps. Cleaning /var/cache/edb and running emerge --metadata does fix the problem, but I wouldn't point my finger at the filesystem. (In reply to comment #7) > Cleaning /var/cache/edb and running emerge --metadata does fix the problem, but > I wouldn't point my finger at the filesystem. Maybe we just need to automatically delete a corrupt cache entry there. I'll check the code. Created attachment 104456 [details]
'Corrupted' /var/cache/edb file
Indeed, clearing out just the corrupted files from /var/cache/edb followed by a sync fixes the problem. I also wouldn't have blamed the filesystem, since I haven't had any other difficulties. Also, redigesting (--digest at least) doesn't seem to recreate the cache item. I've attached the offended mpc-0.12.0 file from /var/cache/edb. There were no disk or read errors when copying it, and nothing else that I could see about it that caused problems. I can also provide more examples from the other machine, if they'd be of use?
*** Bug 158676 has been marked as a duplicate of this bug. *** If someone is able to reproduce this problem, please create run the command with --debug and FEATURES=python-trace enabled. For example: env FEATURES=python-trace emerge --debug mpc >& debug.log tail -n 2000 debug.log > small_debug.log The tail of the log should allow me to see the nature of the exception that is occurring. If anyone experiencing this problem has something in /etc/portage/modules (a non-default cache module) then please be sure to mention that. Created attachment 104471 [details, diff] patch for debugging purposes If you can reproduce the problem with this patch then it should help me see when is causing this. Either output with this patch or a debug log as described in Comment #11 would be quite helpful. Created attachment 104477 [details]
Small debug log from failed mpc installation attempt
Ok, here's a small debugging log. It was from the command:
FEATURES="python-trace" emerge --debug =mpc-0.12.0 >& debug.log
tail -n 2000 debug.log > small_debug.log
The emerge --info is the same as previously mentioned, and there are no modules in /etc/portage/, just sets, bin and postsync.d (which seem only to contain the expected files, nothing unusual in there).
As mentioned, I can still recreate this, so if you need me to apply the debugging patch, or you need any further output, just let me know (although it may take me a few days to get the output, my working hours are all over the place for the next day or two)... 5:)
Thanks! (In reply to comment #14) > Created an attachment (id=104477) [edit] > Small debug log from failed mpc installation attempt > > Ok, here's a small debugging log. It was from the command: Thanks, that was perfect. It's fixed in svn r5342. This has been released in 2.1.2_rc3-r9. *** Bug 158657 has been marked as a duplicate of this bug. *** |