Using beta 2 pre-release, after selecting Mageia 1 install to upgrade, and us keyboard, install fails with * error: panic: sv_setpvn called with negative strlen at /usr/lib/libDrakX/c.pm line 10
Created attachment 1744 [details] report.bug from error
Blocks: (none) => 3342Assignee: bugsquad => thierry.vignaudSource RPM: drakxtools-backend-13.89.2-1.mga2 => drakxtools
Created attachment 1745 [details] report.bug from clean install attempt As with the upgrade, trying to install with custom partitioning, it fails with * error: panic: sv_setpvn called with negative strlen at /usr/lib/libDrakX/c.pm line 10. It returned me to the partitioning choices screen, so I then tried using existing partitions, which then failed with ... * running: mount /dev/loop0 /mnt/mnt/Mageia-2-beta2-i586-DVD -o remount,rw mount: you must specify the filesystem type * error: mount failed
CC: (none) => ennael1, pterjan
Created attachment 1748 [details] show which function failed Sadly we don't log which function failed and in that case the exception message doesn't show it.
Created attachment 1749 [details] runtime patch If you're not comfortable with patching the stage2 on the ISO image with the previous patch then redo the ISO image (see https://wiki.mageia.org/en/Drakx-installer_tips_and_tricks#modifying_the_stage2), you can just try putting this "patch" file on a floppy or on a USB key and restart the install. The install should auto-detect & load this patch from either the floppy or the USB key Then attach the new logs to this bug report
You'll need to start the install with "linux patch" instead of just "linux" or <enter>
Keywords: (none) => NEEDINFO
Created attachment 1756 [details] report.bug The fuction is get_iso_volume_ids. If it's scanning the entire drive for iso files, that would explain the 4 minute delay before the error message shows up. Even if it's only scanning the partition where it was loaded from, that partition has 130G of data on it.
Kernel sees this USB device getting reconnected more than 300 times: usb 1-2: New USB device found, idVendor=15a4, idProduct=1336
Was the original install also from an iso on disk ? I think it only searched for *.iso in the given directory. Is there one which is not in the list of detected ones? avg_arl_cdi_all_100_101201a3281.iso 7100.0.090421-1700_x86fre_client_en-us_retail_ultimate-grc1culfrer_en_dvd.iso mageia-dvd-1-x86_64.iso lupu-511.iso fdfullcd.iso Mageia-2-beta2-i586-DVD.iso ophcrack-xp-livecd-2.3.1.iso mageia-dvd-1-i586.iso boot.iso linuxconsole1.0.2010-cd.iso myiso.iso I would guess it crashes while trying to get the header on another .iso file, so it would not display it
They are all in the same directory. I normally install from the iso on a hard drive, as that's the fastest method. As part of qa testing though, I also test installing from usb, and physical dvd. Other then Mageia iso's, all of the iso images in that directory have been there since at least Feb 2011, so they were there when I installed Magea 1 from that directory. I created a new partition, after shrinking one, and copied the iso to that filesystem, and I can now complete the installation. I don't think the installer should be looking at the other iso images though.
Humm we could either (or both): - make get_iso_volume_ids() more robust (issue might be new since mga1 due to new perl being more sensitive) - not scan for other ISOs (at least not here for what we do with them. See http://svnweb.mageia.org/soft/drakx/trunk/perl-install/c/stuff.xs.pl?view=markup lines 545-555
CC: (none) => jquelin
I'll alter install:media::_look_for_ISO_images() to look only for the known used ISO since we know which one is it and we don't support multi-CD install for a cople years. What's more in your case, in would ignore your ISO since its named myiso.iso (we check for "-(disc|cd|dvd)" It would be nice if you could answer comment #8 though.
Regarding comment 8, all of the iso images in the list are in the same directory the beta 2 iso was loaded from. There is one additional iso in that directory, that is not in the list called ntfsundelete.iso which is from http://ntfsundelete.com/
That site only offers a .exe, do you mean you generated the iso from that .exe?
(In reply to comment #13) > That site only offers a .exe, do you mean you generated the iso from that .exe? Looks like they've changed the site in the last three years. I've uploaded a copy to http://www.ody.ca/~dwhodgins/ntfsundelete.iso It's 5.2M.
(In reply to comment #7) > Kernel sees this USB device getting reconnected more than 300 times: > > usb 1-2: New USB device found, idVendor=15a4, idProduct=1336 That's a flash card reader I was trying to get working to put the bug report onto. It was not connected when I first ran into the bug. It can be ignored as a hardware error unrelated to this bug.
It seems Application id is not set: $ isoinfo -d -i ntfsundelete.iso CD-ROM is in ISO 9660 format System id: Volume id: NTFS_UNDELETE Volume set id: Publisher id: Data preparer id: Application id: Copyright File id: Abstract File id: Bibliographic File id: Volume set size is: 1 Volume set sequence number is: 1 Logical block size is: 2048 Volume size is: 2657 Joliet with UCS level 3 found NO Rock Ridge present The code needs to check its existence before using it.
Fixed in SVN
Status: NEW => RESOLVEDResolution: (none) => FIXED