[Leaplist] Help! My Ubuntu partially freezes up *after* bootup

andrei raevsky raevsky.andrei at gmail.com
Tue Mar 4 15:32:24 EST 2008


Thanks for the pointers!

On Tue, Mar 4, 2008 at 5:50 PM, John Simpson <jms1 at jms1.net> wrote:
> On 2008-03-03, at 1551, andrei raevsky wrote:
>  >
>  > When this happens I can still move the cursor around, but I cannot do
>  > anything else. Here is the output of /var/log/messages when this
>  > happens:
>  >
>  > -------
>
> > Mar  2 08:37:27 Ubuntu2 kernel: [  427.737016]          res
>  > 51/40:08:87:00:88/00:00:00:00:00/e0 Emask 0x9 (media error)
>
> > Mar  2 08:37:39 Ubuntu2 kernel: [  440.095760]          res
>  > 51/40:08:87:00:88/00:00:00:00:00/e0 Emask 0x9 (media error)
>
> > Mar  2 08:37:51 Ubuntu2 kernel: [  452.465642]          res
>  > 51/40:08:87:00:88/00:00:00:00:00/e0 Emask 0x9 (media error)
>
> > Mar  2 08:38:04 Ubuntu2 kernel: [  464.824415]          res
>  > 51/40:08:87:00:88/00:00:00:00:00/e0 Emask 0x9 (media error)
>
> > -------
>  >
>  > This goes on and on and on. The only thing I can do is reboot.
>
>  these message indicate an error with an ATA device, either the hard
>  drive or the CD-ROM. and if you're not trying to read a CD, then it's
>  the hard drive. and the phrase "media error" gives me the impression
>  that the drive is getting ready to go south.
>
>  there may be more specific information within the messages themselves,
>  but i'm not an expert with ATA devices or commands. maybe bryan can
>  shine a brighter light on exactly what these messages are telling you?
>  (i'm pretty sure the number in square brackets, the 427, 440, 452, and
>  464, are timestamps- how many seconds since the kernel started
>  running... but anything after that, i'm lost. maybe it's a physical
>  location on the disk, and there's a bad sector there?)
>
>
>
>  > I would just add that this problem only occurs after bootup anywhere
>  > from 1 to 3 times in a row and that after that everything works fine.
>  > I have no reason to suspect a hardware failure.
>
>  it could be something heat-related... maybe the drive got bumped, and
>  one of the internal components was knocked off of its normal
>  alignment, and it's able to compensate when the system first boots up-
>  but once it's been running for a few minutes and starts to build up a
>  little bit of heat, that pulls it too far out of tolerance... i don't
>  know, i'm just guessing.
>
>
>
>  > I have asked the same question elsewhere, and here is the only reply I
>  > have received:
>  >
>  > -------QUOTE-------
>  > This sounds like it could be your hard drive. Given that you are able
>  > to boot and shutdown, I suggest going to an X terminal; logging in as
>  > root (su); and entering /sbin/shutdown -r -F "now". This will force an
>  > fsck on the next boot.
>
>  if you want to do an fsck, then do an fsck. no need to "trick" the
>  system like this. it sounds like whoever told you this is either a
>  windoze user who's relatively new to linux, because the only way to
>  force windows to run chkdsk on boot is to trick it into doing so, or
>  they don't feel like explaining how to do it by hand, which is:
>
>  assuming you're using grub as the boot loader... when it's booting,
>  press an arrow key up or down to stop the timer, then move to the
>  first option and hit "e"... then, move up or down (if necessary) to
>  the "kernel" line, and hit "e" again. at the end of the line type "
>  single" (i.e. a space, then the word "single") and hit ENTER. then hit
>  "b" to boot. the machine will boot into single-user mode.
>
>  then just type "e2fsck -fy /dev/hda1" (or whatever device contains the
>  filesystem.) the "e2fsck" command is for ext2/ext3 filesystems. the "-
>  f" means to check even if it doesn't "look" like it needs it, and the
>  "-y" means to answer yes to all of the questions.
>
>  if you're using some other filesystem, there should be a similar
>  command with similar options- check the documentation on your system.
>  "fsck" is a generic term, and "e2fsck" is the specific version for
>  ext2/ext3 filesystems. some of the others may have names like
>  "fsck.xfs", but as a rule they all have "fsck" in their names- so "man
>  -k fsck" should give you a list of them (as long as "makewhatis" has
>  been run on the system.)
>
>
>
>  > But why is that be a hard disk problem if the hard disk works
>  > perfectly (except for right after booting up)? do you think that this
>  > is a logical error or a physical one?
>
>  metal expands slightly as it heats up, and when that happens,
>  something inside the drive is being pulled far enough out of place to
>  interfere with data being read? maybe one of the chips on the drive's
>  controller board is getting ready to go, and it's heat sensitive?
>
>  when i've seen cases like this in the past, where a drive works for a
>  certain period of time but then stops, and where turning it off for a
>  while allows it to come back again... it's almost always been heat-
>  related in one way or another. i've seen people do crazy things
>  like... if a drive gets really hot and stops working after a certain
>  length of time, put it into a freezer to keep it cold, and be able to
>  run it for long enough to copy the data off of it.
>
>  again, i'm just guessing here.
>
>
>
>  > Since I never had to use fsck in the past I also wonder whether
>  > answering "yes" to all questions is safe.
>
>  if the problems are logical rather than physical, then answering "yes"
>  to everything will result in your saving as much data as is possible-
>  shy of manually checking and editing the sectors by hand (which
>  requires you to know the low-level details of how the filesystem is
>  stored on the disk. i tried it once, to un-delete a file, and it was
>  some pretty hairy stuff.) anything it can't save is probably something
>  which couldn't have been saved anyway.
>
>  and if the problems are physical rather than logical, it really
>  doesn't matter either way- if the drive is gonna die, it's gonna die.
>
>  i think i would make it my first priority to get a full backup of
>  what's on the drive, before it goes any further south... and THEN try
>  the fsck thing.
>
>  --------------------------------------------------------
>  | John M. Simpson  --  KG4ZOW  --  Programmer At Large |
>  | http://www.jms1.net/                 <jms1 at jms1.net> |
>  --------------------------------------------------------
>  |   Hope for America  --  http://www.ronpaul2008.com/  |
>  --------------------------------------------------------
>
>
>
>
>
>
> _______________________________________________
>  Leaplist mailing list
>  Leaplist at leap-cf.org
>  http://lists.leap-cf.org/mailman/listinfo/leaplist
>
>


More information about the Leaplist mailing list