Help needed fsck exited with signal 11

David Vu David.Vu at
Wed Nov 14 17:57:00 CST 2001

Hi Chris,Matt,

The box has 1GB of RAM which matches your description.  I think I'll
try the mem= workaround now just to get the box going again, but
thanks for the info about your patched kernel.

Hopefully this will enable me to run bonnie++ and to catch those disk
errors.  Crossing fingers now...



: -----Original Message-----
: From: Chris Pascoe [mailto:c.pascoe at]
: Sent: Wednesday, November 14, 2001 10:25 PM
: To: David Vu
: Cc: linux-poweredge at; linux-aacraid-devel at
: Subject: Re: Help needed fsck exited with signal 11
: On Wed, 14 Nov 2001, David Vu wrote (on linux-poweredge):
: > Checking root filesystem
: > / was not cleanly unmounted, check forced
: > Warning... fsck.ext2 for device /dev/sda3 exited with signal 11
: > [FAILED]
: Hi David,
: This (as well as everything being fine with the earlier
: kernels, and on
: rpm -Va) is consistent with problems that I've seen in the RH errata
: kernel on highmem-enabled machines.  See
: linux-aacraid-devel for further
: details.  An unofficial and unsupported patch (written by
: me) was posted
: there yesterday against the RH errata kernel that works
: around the issues
: that become evident with the aacraid driver - but I don't believe it
: fixes the actual cause.
: If you have only 1GB of RAM in the machine, you can work around the
: problem temporarily by sacrificing the top 128MB of RAM -
: add "mem=896M"
: to your LILO line.  Effectively you're disabling highmem by
: doing this,
: which is one of the factors needed for the failure to occur.
: If you want to try a kernel RPM that I've built locally and
: have been
: stressing with continual reboots every 10 minutes for the
: past 24 hours
: (something that would fail 19/20 times for me prior to my
: patching) - then
: you can try:

It's built from the redhat 2.4.9-12 sources, with my workaround added
avoid the crashing I was seeing.  It only comes with a "works for me"
guarantee.  If you want to rebuild from source rather than try this
RPM, then see the aacraid list for the patch - a variant is attached
to a
recent message sent by Matt Domsch.

Christopher Pascoe
Systems Programming Manager
School of Information Technology and Electrical Engineering
The University of Queensland   Brisbane  QLD  4072  Australia

More information about the Linux-PowerEdge mailing list