fsck/usplash: ESC does not kill fsck
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
sysvinit (Ubuntu) |
Fix Released
|
Undecided
|
Martin Pitt | ||
Hardy |
Fix Released
|
Undecided
|
Martin Pitt |
Bug Description
Binary package hint: sysvinit
While similar to bug 194017, this bug is a result of "fsck" not getting killed correctly. After capturing $FSCKPID with $! (instead of using job control %1), running "ps" during various stages of the script shows that "logsave" is killed, but that fsck continues running:
killing (3744)...
root 3744 0.0 0.1 3844 564 ? S 23:57 0:00 logsave -s /var/log/
root 3745 0.0 0.1 18816 944 ? S 23:57 0:00 fsck -C3 -R -A -a
root 3812 41.0 0.2 23360 1304 ? R 23:57 0:00 fsck.ext3 -a -C3 /dev/fakedev2
root 3744 0.0 0.1 3844 564 ? S 23:57 0:00 logsave -s /var/log/
sent kill...
root 3745 0.0 0.1 18816 944 ? S 23:57 0:00 fsck -C3 -R -A -a
root 3812 43.0 0.2 23360 1304 ? R 23:57 0:00 fsck.ext3 -a -C3 /dev/fakedev2
root 3826 0.0 0.1 3944 608 ? R 23:57 0:00 grep fsck
sleeping...
root 3745 0.0 0.1 18816 944 ? S 23:57 0:00 fsck -C3 -R -A -a
root 3812 45.0 0.2 23360 1304 ? R 23:57 0:00 fsck.ext3 -a -C3 /dev/fakedev2
waiting (3744)...
root 3745 0.0 0.1 18816 944 ? S 23:57 0:00 fsck -C3 -R -A -a
root 3812 53.0 0.2 23360 1304 ? R 23:57 0:03 fsck.ext3 -a -C3 /dev/fakedev2
root 3841 0.0 0.1 3944 612 ? R 23:58 0:00 grep 3744
waited (exit code 143)...
root 3745 0.0 0.1 18816 944 ? S 23:57 0:00 fsck -C3 -R -A -a
root 3812 53.6 0.2 23360 1304 ? R 23:57 0:03 fsck.ext3 -a -C3 /dev/fakedev2
Related branches
Changed in sysvinit: | |
assignee: | nobody → pitti |
milestone: | none → ubuntu-8.04 |
Changed in sysvinit: | |
status: | New → In Progress |
Changed in sysvinit: | |
milestone: | ubuntu-8.04 → none |
Changed in sysvinit: | |
status: | In Progress → Fix Committed |
This bug was fixed in the package sysvinit - 2.86.ds1- 14.1ubuntu42
--------------- 14.1ubuntu42) hardy; urgency=low
sysvinit (2.86.ds1-
* debian/ initscripts/ lib/init/ usplash- fsck-functions. sh:
- When pressing ESC, do not kill the logsave command (since this does not
reliably kill the subprocess), but the fsck.ext[23] processes directly.
(LP: #203322)
- When checking multiple partitions in succession, print the name of the
next device when it comes to its turn. NB that this still does not work
with checking multiple partitions in parallel, that's an fsck bug.
(LP: #203323)
- Use usplash 0.5.17's new FAIL_NO_USPLASH mode to avoid being stuck at
reading from the usplash FIFO if usplash died during fsck.
- Also output fsck progress information to the current console. This looks
quite dodgy ATM, but is better than not printing anything at all if
usplash died in between.
-- Martin Pitt <email address hidden> Wed, 26 Mar 2008 13:35:10 +0100