Re: Disk errors?
Christopher Judd <email@example.com> writes:
> On Monday 14 June 2010 15:18:50 Christopher Judd wrote:
>> On Monday 14 June 2010 14:48:13 Ron Johnson wrote:
>> > On 06/14/2010 01:00 PM, brian m. carlson wrote:
>> > > On Mon, Jun 14, 2010 at 01:23:07PM -0400, Christopher Judd wrote:
>> > [snip]
>> > > You probably don't want to put that on a public mailing list. It looks
>> > > silly.
>> > Silly or not, he probably has to do it.
>> > My company used to mandate such a disclaimer, but then we were
>> > bought out by a Big Co that must realize who stupid the disclaimer is.
>> Exactly. It is added to all outgoing mail by the server here; I have no
>> control over it.
>> It turned out that my root partition, /dev/sda1, was marked unclean
>> (other partitions were OK. After booting from a rescue disk and running
>> e2fsck, it booted normally with no errors yet (up about 40 min). Shouldn't
>> that be detected during the boot process and fsck run before mounting it?
> I guess that this can't be done, since fsck.ext2 resides in /sbin, and
> wouldn't be available prior to mounting /.
Which is why / is first mounted read-only and then fsck is run on /. If
fsck makes any changes to the filesystem the system reboots to avoid
conflicts between cached blocks and blocks fsck changed on disk.
But I think the problem is that fsck is currently broken. It never
checks anything but just prints the version string and exits. At least I
noticed that recently but haven't debugged it yet.