Description of problem: On real hardware: Clean installed M3.1 and updated no package. No changes to grub, only ensured systemctl showed working ip(6)tables and shorewall(6) services. Upon finishing installs of packages: An error occurred: ERROR: 'script' failed for mageia-gfxboot-theme=4.4.5.12-1.mga4.x86_64: and next: An error occurred mkinitrd failed (mkinitrd -v -f /boot/initrd-3.12.3-desktop-1.mga4.img 3.12.3-desktop-1.mga4). | Ok | #every time OK is pressed it happens again From tty can be seen ...at /usr/lib/libDrakX/bootloader.pm line 137 which is from drakxtools-backend-16.12 Current Cauldron has drakxtools-ackend-16.13, but I could not use it as workaround, because the fs at /usr/lib/libDrakX is read only, so I couldn;t copy bootloader.pm form drakxtools-backend-16.13 to there from under current cauldron in another mounted partition Will attach report.bug.xz Hardware is described as entry #3 at: https://wiki.mageia.org/en/QA_iso_hardware_list#Notebooks Reproducible: Steps to Reproduce:
Created attachment 4605 [details] report from /root/drakx
To be read as (inserted) line 7 in my report: quote Upgrading such M3.1 with this iso: unquote -- Sorry I missed it.
Blocks: (none) => 11827
CC: (none) => ennael1, marja11Assignee: bugsquad => thierry.vignaudWhiteboard: (none) => 4beta2
CC: (none) => davidwhodgins, eeeemail
CC: (none) => mageia
Blocks: (none) => 11979
Blocks: 11827 => (none)
(In reply to Dick Gevers from comment #1) > Created attachment 4605 [details] > report from /root/drakx Are you sure this is correct? It only mentions mkinitrd once and for an mga3 kernel. Looking at the package names it all appears to be mga3 packages, not anything related to mga4...
That's a dracut issue. We can't do anything if mkinitrd fails. BTW, I'm not sure the attached report.bug.xz matches this issue as I fail to see such failure beyond those errors which seems to have been happily ignored by dracut: find: `standard output': Broken pipe find: write error (did you've a full fs on this one?) It also miss the script error your reported
Assignee: thierry.vignaud => bugsquadSource RPM: Mageia-4-beta2-x86_64-DVD.iso round 4 => dracut
Comment on attachment 4605 [details] report from /root/drakx @Colin: good catch! I hadn't seen this! Kernel version indeed is 3.8.13 and this is DrakX v15.54...
Attachment 4605 is obsolete: 0 => 1CC: (none) => thierry.vignaud
Adding NEEDINFO. Also we got random mkinitrd failures for quite a long time, so I remove this one from the tracker
Keywords: (none) => NEEDINFOBlocks: 11979 => (none)
Blast :(( So I found the wrong report. But it was the correct DVD, and presently the same /mnt/beta/root/drakx exists per 'stat' even though I reproduced diskdrake detaching from installer. Does /root/drakx get written before the install is finished? Where?
Keywords: NEEDINFO => (none)
(In reply to Dick Gevers from comment #7) > Blast :(( > > So I found the wrong report. But it was the correct DVD, and presently the > same /mnt/beta/root/drakx exists per 'stat' even though I reproduced > diskdrake detaching from installer. > > Does /root/drakx get written before the install is finished? Where? You attached report.bug.xz, didn't you? After all my 4beta2 upgrades, the report.bug.xz in /root/drakx/ was of the original install, not of the upgrade. Also, in all those cases, there was a report.bug of the upgrade in /root/drakx/
Comment on attachment 4605 [details] report from /root/drakx trying hard to report ddebug.log from same folder w/o success till now
Created attachment 4609 [details] ddebug.log ref. upgrade from /root/drakx
Comment on attachment 4609 [details] ddebug.log ref. upgrade from /root/drakx No mkinitrd error but looks corrupted
Attachment 4609 is obsolete: 0 => 1
This is all I've got when upgrading and unable to finish same (plus auto_inst.cfg.pl). Please (anyone) describe how to record/report such upgrade pecisely, otherwise a bug report becomes impossible (which might be a next bug ;)) Thanks in advance.
Can you reproduce the mkinitrd bug by running it manually after it's failed as part of the general upgrade? If so the output from that would be useful (you can also run dracut -vvv manually too which should show you more info about why it's failing).
And also could you answer about the space issue (near full fs?)?
@coling re #c13 I will need to reinstall M3.1 and try. Frankly I would have thought that I would have done so myself, but not sure if I tried or not. @tv re #c14 I figured from earlier probs that 34M was enough for a /boot partition. But this proves it is not when upgrading, 'cause one image, an "old" image and a new one is nearly 30 M. Will try to fix this before I go for reproduction.
As such I find my bugreport is invalid, because with a larger partition /boot mkinitrd does not fail. But on the other hand, when upgrading, I believe the installer should look out for possible bottlenecks if the box could get out of ample space when upgrading. And that is what was at the basis of this error: I fixed my partition table and removed previously used partition for / <alpha |beta |rc> /boot at 38 MB and created another elsewhere on the HD with a size of 150 MB, so failure did not occur anymore. Please advise if there should be a bugreport for such? BTW, Moreover I noted that diskdrake under certain circumstances does create overlapping partitions and errs out afterwards and also disables gparted from accessing the HD after the layout has become 'overlapping'. Will report this separately. This was why I it took me so long to comment again (plus personal other detractions).
Keywords: (none) => NEEDINFO
See also https://bugs.mageia.org/show_bug.cgi?id=12012
Status: NEW => RESOLVEDResolution: (none) => INVALID