Red Hat Bugzilla – Bug 129968
test kernel 2.6.8-1.520 makes umount to go into D state
Last modified: 2015-01-04 17:08:46 EST
Description of problem:
While remaking some initrd image the whole procedure ends in
an unkillable umount like this:
root 24885 0.0 0.0 41696 452 pts/0 D 17:39 0:00
\_ umount /tmp/initrd.mnt.o24736
This is part of ps output from some five minutes later after
various failed attempts to dislodge the whole thing. Even
with an emergency sync and r/o reboot via sysrq I ended up with
a damage on /boot file system. Also attempts to shutdown orderly
failed and I had to use sysrq "reboot" to get restore an order.
A processor state and tasks from sysrq are attached; but it looks
like this is an NFS code again even if I did not try to unmount
an NFS mount.
Version-Release number of selected component (if applicable):
I would rather not try. :-)
Created attachment 102754 [details]
data collected via sysrq after umount got into D state
I meant, of course, "r/o remount via sysrq" and not "reboot". Sigh!
mass update for old bugs:
Is this still a problem with the 2.6.9 based update kernel ?
> Is this still a problem with the 2.6.9 based update kernel ?
I do not think so. I did not see anything like that for a long time.