Red Hat Bugzilla – Bug 1252465
rescue mode crashes when trying to rescue a dual-boot system
Last modified: 2015-08-11 13:46:47 EDT
Created attachment 1061527 [details] anaconda traceback Description of problem: I have installed a dual-boot system with Windows 8 and Fedora 23 Alpha, in UEFI mode. When I boot F23 Alpha Workstation netinst and select "Rescue system", I end up in a tmux console with anaconda traceback. The error is in fsminsize.py:57 and the error is: > UnboundLocalError: local variable 'e' referenced before assignment The error is not logged anywhere, and no anaconda-tb file is created, so there seems to be a second bug or an omission in logging and crash handling. As a punishment, all you get is a camera shot, because I don't know how else to get the info. This problem seems to be specific to my disk contents. I have tried the same with a clean F23 Alpha install, and that seems to work. So this is likely related to the dual boot setup. In the log I see some output related to NTFS partitions: 14:59:34,187 INFO program: Running... ntfsinfo -m /dev/sda4 14:59:34,198 INFO program: stdout: 14:59:34,198 INFO program: Failed to open '/dev/sda4'. 14:59:34,198 INFO program: stderr: 14:59:34,199 INFO program: b'Volume is scheduled for check.' 14:59:34,200 INFO program: b"Please boot into Windows TWICE, or use the 'force' option." 14:59:34,200 INFO program: b'NOTE: If you had not scheduled check and last time accessed this volume' 14:59:34,200 INFO program: b'using ntfsmount and shutdown system properly, then init scripts in your' 14:59:34,200 INFO program: b'distribution are broken. Please report to your distribution developers' 14:59:34,201 INFO program: b'(NOT to us!) that init scripts kill ntfsmount or mount.ntfs-fuse during' 14:59:34,201 INFO program: b'shutdown instead of proper umount.' 14:59:34,201 DEBUG program: Return code: 1 14:59:34,202 INFO program: Running... ntfsresize -m /dev/sda4 14:59:34,212 INFO program: stdout: 14:59:34,212 INFO program: ntfsresize v2015.3.14 (libntfs-3g) 14:59:34,213 INFO program: ERROR: Volume is scheduled for check. 14:59:34,213 INFO program: Run chkdsk /f and please try again, or see option -f. 14:59:34,213 DEBUG program: Return code: 1 I tried booting into Windows twice in a row, that didn't help. Maybe it would help if I forced filesystem check from windows. But anyway, anaconda should not crash, even if this was something that could not be fixed from Linux. Version-Release number of selected component (if applicable): F23 Alpha anaconda 23.17-1 How reproducible: always for my particular disk layout Steps to Reproduce: 1. install Windows 8 and Fedora 23 Alpha in dual-boot, UEFI mode 2. boot F23 Alpha Workstation netinst (UEFI or BIOS mode) 3. see crash immediately on start
Created attachment 1061528 [details] journal
Created attachment 1061529 [details] anaconda.log
Created attachment 1061530 [details] program.log
Created attachment 1061531 [details] storage.log
Created attachment 1061532 [details] syslog
This seems to violate: "The rescue mode of the installer must start successfully and be able to detect and mount any installation performed according to the applicable criteria, and provide a shell with access to utilities capable of performing typical recovery operations. " https://fedoraproject.org/wiki/Fedora_23_Beta_Release_Criteria#rescue-mode A dual-boot system is a pretty common system that people might want to perform a rescue operation on.
Fixed in the newest blivet. *** This bug has been marked as a duplicate of bug 1249304 ***