Toshiba Satellite running KDE

[Meemaw]

My daughter updated her laptop about Oct 5th... the battery ran down over the weekend so she plugged it in and restarted it... now it won't boot at all... keeps getting a fs error...    if you run fsck you get 'block 6651906 has errors'     ignore?     force rewrite? if she types startx, it says /dev/sda1 is missing (but I think that's the hard drive....   and I think she gave away her last live cd.....)

I'm sure this isn't enough information.  Any ideas anyway?

 

[Just17]

My daughter updated her laptop about Oct 5th... the battery ran down over the weekend so she plugged it in and restarted it... now it won't boot at all... keeps getting a fs error...    if you run fsck you get 'block 6651906 has errors'     ignore?     force rewrite? if she types startx, it says /dev/sda1 is missing (but I think that's the hard drive....   and I think she gave away her last live cd.....)

I'm sure this isn't enough information.  Any ideas anyway?

 

get another LiveCD and run fsck from that to repair the fs 

[Meemaw]

My daughter updated her laptop about Oct 5th... the battery ran down over the weekend so she plugged it in and restarted it... now it won't boot at all... keeps getting a fs error...    if you run fsck you get 'block 6651906 has errors'     ignore?     force rewrite? if she types startx, it says /dev/sda1 is missing (but I think that's the hard drive....   and I think she gave away her last live cd.....)

I'm sure this isn't enough information.  Any ideas anyway?

 

get another LiveCD and run fsck from that to repair the fs 

We can try that, but the problem is that she lives 50 miles away... and when I'm in a terminal, I turn into an idiot....  I'll see what I can figure out...    thanks!
 

[Just17]

My daughter updated her laptop about Oct 5th... the battery ran down over the weekend so she plugged it in and restarted it... now it won't boot at all... keeps getting a fs error...    if you run fsck you get 'block 6651906 has errors'     ignore?     force rewrite? if she types startx, it says /dev/sda1 is missing (but I think that's the hard drive....   and I think she gave away her last live cd.....)

I'm sure this isn't enough information.  Any ideas anyway?

 

get another LiveCD and run fsck from that to repair the fs 

We can try that, but the problem is that she lives 50 miles away... and when I'm in a terminal, I turn into an idiot....  I'll see what I can figure out...    thanks!
 

Just make sure the partitions being checked are not mounted .....  maybe run as root

Code: [Select]

umount -a
before using

fsck

Good luck 

[Meemaw]

So she's running fsck from a live cd... what does she do when she gets an error message? Ignore or what? We possibly didn't put in the correct options to get us what we want......

(sorry, we're way out of our level of competence here.....)

 

[Just17]

So she's running fsck from a live cd... what does she do when she gets an error message? Ignore or what? We possibly didn't put in the correct options to get us what we want......

(sorry, we're way out of our level of competence here.....)

 

She should get an offer from fsck to repair the problems found.

If she gets a warning about the partition being mounted she should abort immediately and unmount the partition before running fsck.

Here is the result from a chack I did on a partition of a USB stick.
As you can see it found it was incorrectly unmounted so it did its checks.


fsck /dev/sdm1

Code: [Select]

fsck from util-linux-ng 2.18
e2fsck 1.41.14 (22-Dec-2010)
LivePCLOS was not cleanly unmounted, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
LivePCLOS: 90/252960 files (0.0% non-contiguous), 183883/1010432 blocks

The options   -p   or   -y    could be used .....  I prefer   -y    so that I can see what it is doing and yet not have to say yes to everything. Using   -p   will hide the output I think.

So I would use

fsck -y /dev/sdm1

for the partition above.


regards

[Old-Polack]

So she's running fsck from a live cd... what does she do when she gets an error message? Ignore or what? We possibly didn't put in the correct options to get us what we want......

(sorry, we're way out of our level of competence here.....)

 

She should get an offer from fsck to repair the problems found.

If she gets a warning about the partition being mounted she should abort immediately and unmount the partition before running fsck.

Here is the result from a chack I did on a partition of a USB stick.
As you can see it found it was incorrectly unmounted so it did its checks.


fsck /dev/sdm1

Code: [Select]

fsck from util-linux-ng 2.18
e2fsck 1.41.14 (22-Dec-2010)
LivePCLOS was not cleanly unmounted, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
LivePCLOS: 90/252960 files (0.0% non-contiguous), 183883/1010432 blocks

The options   -p   or   -y    could be used .....  I prefer   -y    so that I can see what it is doing and yet not have to say yes to everything. Using   -p   will hide the output I think.

So I would use

fsck -y /dev/sdm1

for the partition above.


regards

Better fsck -fy /dev/sdm1.

The -f forces a complete inode by inode check, rather than just a journal check. That is what we see in the message from the unclean umount;  check forced.

[Just17]

So she's running fsck from a live cd... what does she do when she gets an error message? Ignore or what? We possibly didn't put in the correct options to get us what we want......

(sorry, we're way out of our level of competence here.....)

 

She should get an offer from fsck to repair the problems found.

If she gets a warning about the partition being mounted she should abort immediately and unmount the partition before running fsck.

Here is the result from a chack I did on a partition of a USB stick.
As you can see it found it was incorrectly unmounted so it did its checks.


fsck /dev/sdm1

Code: [Select]

fsck from util-linux-ng 2.18
e2fsck 1.41.14 (22-Dec-2010)
LivePCLOS was not cleanly unmounted, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
LivePCLOS: 90/252960 files (0.0% non-contiguous), 183883/1010432 blocks

The options   -p   or   -y    could be used .....  I prefer   -y    so that I can see what it is doing and yet not have to say yes to everything. Using   -p   will hide the output I think.

So I would use

fsck -y /dev/sdm1

for the partition above.


regards

Better fsck -fy /dev/sdm1.

The -f forces a complete inode by inode check, rather than just a journal check. That is what we see in the message from the unclean umount;  check forced.

Wouldn't do any harm I guess ......  but the 'f' option is for situations where the partition is marked clean .....  which is not the situation I believe.

Personally I would not use the 'f' option until and unless I had feedback from running without it first.

regards.

[Meemaw]

She should get an offer from fsck to repair the problems found.

If she gets a warning about the partition being mounted she should abort immediately and unmount the partition before running fsck.

Here is the result from a chack I did on a partition of a USB stick.
As you can see it found it was incorrectly unmounted so it did its checks.


fsck /dev/sdm1

Code: [Select]

fsck from util-linux-ng 2.18
e2fsck 1.41.14 (22-Dec-2010)
LivePCLOS was not cleanly unmounted, check forced.
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
LivePCLOS: 90/252960 files (0.0% non-contiguous), 183883/1010432 blocks

The options   -p   or   -y    could be used .....  I prefer   -y    so that I can see what it is doing and yet not have to say yes to everything. Using   -p   will hide the output I think.

So I would use

fsck -y /dev/sdm1

for the partition above.


regards

Better fsck -fy /dev/sdm1.

The -f forces a complete inode by inode check, rather than just a journal check. That is what we see in the message from the unclean umount;  check forced.

Wouldn't do any harm I guess ......  but the 'f' option is for situations where the partition is marked clean .....  which is not the situation I believe.

Personally I would not use the 'f' option until and unless I had feedback from running without it first.

regards.

I think the command we used was fsck /dev/sda1

The result was;
Error reading block 6651906 (Attempt to read block from filesystem resulted in short read) while getting next inode from scan. Ignore error <y>?

Unless she shortened it (she was texting me) that's the output.....

So she would find out more from

fsck -fy /dev/sda1      ?

Thanks bunches to you both!!!!   

(I'm thinking this would make a good mag article too, but I don't think I'm smart / experienced enough to write it!!!!)   

[Just17]

I think the command we used was fsck /dev/sda1

The result was;
Error reading block 6651906 (Attempt to read block from filesystem resulted in short read) while getting next inode from scan. Ignore error <y>?

Unless she shortened it (she was texting me) that's the output.....

So she would find out more from

fsck -fy /dev/sda1      ?

Thanks bunches to you both!!!!  

(I'm thinking this would make a good mag article too, but I don't think I'm smart / experienced enough to write it!!!!)  

She is required to press 'Y' at each question using that command.

By inserting  -y  in the command, it presumes that yes is the answer to all its questions to correct the fault or not.

So I would suggest she use

fsck -y /dev/sda1

and if a verbose output is required for someone to look over later then

fsck -vy /dev/sda1

and the output can be copy/pasted from the terminal.

Also look at the -p  option ......  to automatically repair all errors.

regards

[Meemaw]

I think the command we used was fsck /dev/sda1

The result was;
Error reading block 6651906 (Attempt to read block from filesystem resulted in short read) while getting next inode from scan. Ignore error <y>?

Unless she shortened it (she was texting me) that's the output.....

So she would find out more from

fsck -fy /dev/sda1      ?

Thanks bunches to you both!!!!   

(I'm thinking this would make a good mag article too, but I don't think I'm smart / experienced enough to write it!!!!) 

She is required to press 'Y' at each question using that command.

By inserting  -y  in the command, it presumes that yes is the answer to all its questions to correct the fault or not.

So I would suggest she use

fsck -y /dev/sda1

and if a verbose output is required for someone to look over later then

fsck -vy /dev/sda1

and the output can be copy/pasted from the terminal.

regards

Ooopps!!!   I guess I didn't mean y, then.... I was thinking that she would want to be able to see what was going on rather than having fsck just proceed... so maybe just fsck -v /dev/sda1 so she would have output but still be asked what to do?

And then she is asked whether she should ignore the error...  if she says no, is she then asked if a repair should be attempted?  I'm just trying to find out some steps to give her.....

(I've been looking for a good how-to that doesn't just give you the terminology and then wish you luck.... but one that actually tells you things you can do in different situations.....)

Thanks so much!!!
 

[Just17]

IMO, she should repair everything fsck comes across ........  unless she has the training and knowledge to make decisions about questions like the one asked.

Presently the partition is corrupted, and fsck is the means to fix it - hopefully.

I doubt she could make better decisions than fsck .......  but maybe I am wronging her   

In your position I would give her the  -p  option, and just give fsck its head .......  no questions, no answers .....  it just gets fixed, if possible.

You might want to await another opinion though 


regards

[AS]

Adding my opinion here 

I would run fsck -f -y /dev/sda1

-f option will force the filesystem check even if it's recognized as clean.

-y will answer yes to all questions, from man page

       -y     For some filesystem-specific checkers, the -y option will cause the fs-specific fsck to  always  attempt  to  fix  any
              detected filesystem corruption automatically.  Sometimes an expert may be able to do better driving the fsck manually.

-p option, from manual page:

       -p     Automatically  repair  ("preen")  the  file system.  This option will cause e2fsck to automatically fix any filesystem
              problems that can be safely fixed without human intervention.  If e2fsck discovers a problem  which  may  require  the
              system  administrator  to  take  additional corrective action, e2fsck will print a description of the problem and then
              exit with the value 4 logically or'ed into the exit code.  (See the EXIT CODE section.)  This option is normally  used
              by the system's boot scripts.

Very probably the system already ran the fsck -p at boot ... ultimately prompting for user input.

Additionally I would point the attention to the error:

Error reading block 6651906 (Attempt to read block from filesystem resulted in short read) while getting next inode from scan. Ignore error <y>?

The meaning of this error is that fsck is unable to read one specific block of the hard disk, and it is issued in two cases:
- the block number is wrong, outside the physical range of available disk sectors or outside the partition extent.
- the block is damaged, ultimately this is an hardware error, meaning an hard disk failure, in this specific case, usually you should also see some message logged on console: sda1: I/O Error or something like that.

Note that answer <y> to such question will always lead to loss of some data, (this is however necessary from a filesystem point of view to restore the filesystem consistency). The -y option will fix the filesystem looking from the filesystem perspective, (not from the user perspective).

In case of 'damaged block' I would suggest to backup the important data immediately after successfully cleaned up the filesystem from fsck, also consider additional disk tests and/or disk replacement. Not delete/overwrite previous backup(s), if available.

AS

[Old-Polack]

Meemaw:

From man e2fsck:

-f     Force checking even if the file system seems clean.

-p     Automatically  repair ("preen") the file system.  This option will cause e2fsck to auto‐
              matically fix any filesystem problems that can be safely fixed without  human  interven‐
              tion.   If e2fsck discovers a problem which may require the system administrator to take
              additional corrective action, e2fsck will print a description of the  problem  and  then
              exit  with the value 4 logically or'ed into the exit code.  (See the EXIT CODE section.)
              This option is normally used by the system's boot scripts.  It may not be  specified  at
              the same time as the -n or -y options.

-v     Verbose mode.

-y     Assume an answer of `yes' to all questions; allows e2fsck to be used  non-interactively.
              This option may not be specified at the same time as the -n or -p options.

When all else fails, e2fsck -fy /dev/sdxn will fix the filesystem, if it can be fixed at all. As I've had filesystem checks that failed, with other options, and were then subsequently always successful when the -fy options were specified, they are now always my first choice of options, to save time and effort. If the filesystem cannot be repaired with these options, your only alternative is to reformat, using mke2fs using the the -c option.

From man mke2fs:

-c     Check the device for bad blocks before creating the file  system.   If  this  option  is
              specified twice, then a slower read-write test is used instead of a fast read-only test.

[coyoteawooo]

Right...
Totally...
I'm even farther out of my depth than MeeMaw, so the question really is:
What on earth do I do now?
It doesn't give me the option to answer any question except whether to ignore the error in block 6651906, and ignoring an error just doesn't seem like a good idea to me. I just want it fixeded and haven't the foggiest how one fixeses this problem (hence the phone call to MeeMaw).

I assume that what y'all are trying to say is that I need to give the 'puter a command to "fix yourself, idget!"  Which command would that be?  Or am I here too soon and you haven't quite decided among yourselves?

We did try "fsck /media/sda1" and it found that block error.  I thought maybe "fsck -l /media/sda1" so it could list my bad blocks and I could maybe-might-possibly find out how extensive the damage is (if it's a big ol' chunk of my computer, I can just assume the gods have deemed it poor judgement on my part to have internet installed in my home and turn the ethernet off, thus saving myself the cash and paying penance in the form of computerlessness until such time as i can afford a replacement - one never questions the gods.)

Or, if y'all would prefer, I can put try "fsck -p /media/sda1" to make it try to fix itself without asking me anything it doesn't absolutely have to. Or "fsck -y /media/sda1" so it'll ignore that error I'm afraid to and go on with its bad self.  That's kind of what I've got it narrowed down to, I'm just not smart enough to make the decision on my own.

Thoughts?  Thoughts that don't include unmounting "root," which seems like a horrible-tasting idea? (Pretty sure the 'puter is smart enough not to let me try to do that anyway.)




Oh, and old-polack, that's totally what the manual says.  Yep.  Read that at http://linux.die.net/man/8/fsck.  And while I speak english, bad english, spanglish, male, drunk, drunken male, and several variations thereon, computer manual is not in my list of language subroutines (right up there with tact and human pair bonding).  So while the manual totally says that, I know more latin than manual and it's greek to me.