Bizarre cascade of problems.

Bizarre cascade of problems.

Post by Jason Kir » Sat, 28 Apr 2001 05:54:33



I'm in a bit of a pickle with RedHat 7.0. I installed a new version of
librep using RPM --force because it was telling me the newer version was
actually older (the new version was 0.13.6-1 and the existing one was
0.12.1-3). However the s**t has fit the fan and I was obviously wrong to
do this. It gets better. My boot disc seems to be corrupted and for some
unknown reason the dual boot Windows has been corrupted at the same time
(I'm at a loss to explain that one).

Can any body offer me any advice on how to updo this damage. If I can get
a working boot disc will I just be able to install the original librep
rpm? I'm inclinded to format the drive and start again, but I need at
least one working OS to pull off the stuff since the last back up.

Any advice wil be welcome.

Cheers.

-Jason

--------------------------------------------------------------------------------
                    WWW: www.geocities.com/jason.kirk  
--------------------------------------------------------------------------------
         "No try! Do.. or do not. There is no try." - Yoda, SW: ESB

 
 
 

Bizarre cascade of problems.

Post by Davi » Sat, 28 Apr 2001 08:07:26



> I'm in a bit of a pickle with RedHat 7.0. I installed a new version of
> librep using RPM --force because it was telling me the newer version was
> actually older (the new version was 0.13.6-1 and the existing one was
> 0.12.1-3). However the s**t has fit the fan and I was obviously wrong to
> do this. It gets better. My boot disc seems to be corrupted and for some
> unknown reason the dual boot Windows has been corrupted at the same time
> (I'm at a loss to explain that one).

> Can any body offer me any advice on how to updo this damage. If I can get
> a working boot disc will I just be able to install the original librep
> rpm? I'm inclinded to format the drive and start again, but I need at
> least one working OS to pull off the stuff since the last back up.

> Any advice wil be welcome.

You should be able to use the CD to boot the system.
Boot from CD
At boot prompt enter:

  linux rescue root=/dev/hdaX
                      where X is the correct partition.

This should work depending on if the file system is corrupted.
Mileage may vary.
--
Confucius say: He who play in root, eventually kill tree.
Registered with the Linux Counter.  http://counter.li.org
ID # 123538
Completed more W/U's than 99.177% of seti users. +/- 0.01%

 
 
 

Bizarre cascade of problems.

Post by Jason Kir » Sun, 29 Apr 2001 03:55:45



> This should work depending on if the file system is corrupted.
> Mileage may vary.

Thanks for the advice. I had a look around and decided that the problem
was little beyond me so I decided to reinstall (keeping the /home)
parition. However the installer keeps crashing when it tries to format /
filesystem. I suspect there may be a back block on the partition.

I'm using the Disc Doctor option in the RH7.0 installer and I tell it to
check for bad blocks. Is there any utilities where I can format the
partition before installation that will also fix back sectors corrects?

Cheers,

-Jason Kirk

--------------------------------------------------------------------------------
                    WWW: www.geocities.com/jason.kirk  
--------------------------------------------------------------------------------
         "No try! Do.. or do not. There is no try." - Yoda, SW: ESB

 
 
 

Bizarre cascade of problems.

Post by Davi » Sun, 29 Apr 2001 04:12:17




> > This should work depending on if the file system is corrupted.
> > Mileage may vary.

> Thanks for the advice. I had a look around and decided that the problem
> was little beyond me so I decided to reinstall (keeping the /home)
> parition. However the installer keeps crashing when it tries to format /
> filesystem. I suspect there may be a back block on the partition.

> I'm using the Disc Doctor option in the RH7.0 installer and I tell it to
> check for bad blocks. Is there any utilities where I can format the
> partition before installation that will also fix back sectors corrects?

> Cheers,

> -Jason Kirk

You can try:

   "cat /dev/zero >/dev/hdaX"

where hdaX is the partition you want to format. This will wipe the
partition with "zero" then use mkfs.ext2 or mke2fs to format it back to
ext2 format.

      man mkfs
      man mke2fs

You might also read "man badblocks" if you want to see what it may find.

--
Confucius say: He who play in root, eventually kill tree.
Registered with the Linux Counter.  http://counter.li.org
ID # 123538
Completed more W/U's than 99.177% of seti users. +/- 0.01%

 
 
 

Bizarre cascade of problems.

Post by Jason Kir » Mon, 30 Apr 2001 02:01:20



>    "cat /dev/zero >/dev/hdaX"

> where hdaX is the partition you want to format. This will wipe the
> partition with "zero" then use mkfs.ext2 or mke2fs to format it back to
> ext2 format.

>       man mkfs
>       man mke2fs

> You might also read "man badblocks" if you want to see what it may find.

Tried the cat command, but it just told me "error writing output No Space
left on device" which was helpfull so I decided to try a different
Linux's installer to see if the same thing happened (Mandrake 7.0) and
crashed out at the same stage reporting:

VPS: Cannot open root device on 08:01
Kernel panic VPS Unable to mount root fs on 08:01

All of this worked identically before (I know I kept a log book of the
options I used to install) so I'm beginning to think its a more
fundimental problem with the hard disc.  

What a way to spend a weekend.

-Jason Kirk

--------------------------------------------------------------------------------
                    WWW: www.geocities.com/jason.kirk  
--------------------------------------------------------------------------------
         "No try! Do.. or do not. There is no try." - Yoda, SW: ESB

 
 
 

Bizarre cascade of problems.

Post by Davi » Mon, 30 Apr 2001 03:35:22




> >    "cat /dev/zero >/dev/hdaX"

> > where hdaX is the partition you want to format. This will wipe the
> > partition with "zero" then use mkfs.ext2 or mke2fs to format it back to
> > ext2 format.

> >       man mkfs
> >       man mke2fs

> > You might also read "man badblocks" if you want to see what it may find.

> Tried the cat command, but it just told me "error writing output No Space
> left on device" which was helpfull so I decided to try a different
> Linux's installer to see if the same thing happened (Mandrake 7.0) and
> crashed out at the same stage reporting:

> VPS: Cannot open root device on 08:01
> Kernel panic VPS Unable to mount root fs on 08:01

> All of this worked identically before (I know I kept a log book of the
> options I used to install) so I'm beginning to think its a more
> fundimental problem with the hard disc.

If you used the "cat" command above then you will still need to format
the partition back to the "ext2" format since it just wiped the
partition with  zero's or you might even use a DOS/windoz boot disk
after the "cat" command to fix the MBR with  "fdisk /MBR".

--
Confucius say: He who play in root, eventually kill tree.
Registered with the Linux Counter.  http://counter.li.org
ID # 123538
Completed more W/U's than 99.177% of seti users. +/- 0.01%

 
 
 

1. RH Linux: Bizarre wtmp/utmp, log file, and file date problems

In all my years of UNIX sys admin experience I have never seen something
so bizarre as this happen, first seen this morning on my Red Hat 5.2
Linux system:

- New logins are no longer being reported with 'w', 'who', and 'last'
commands.  Entries are being made to utmp/wtmp but I haven't verified if
they're correct.  Rotating/removing these files doesn't help either.

- /var/log/secure is no longer being logged to (scary).

- The system date as reported by 'date' and last mod dates on new files
on all filesystems differ by exactly 5 hours, the file being 5 hours in
the future.  Maybe this has something to do with EST timezone?

- The MAIL env var is being set to /spool/mail/login instead of the
proper /var/spool/mail/login.  This causes mail apps to fail to find the
users incoming mail spool.  Resetting this variable manually fixes
things.

Things I've tried doing to fix the problems:

- Rebooting the system (but not power cycling, it's remote).

- Messing with wtmp/utmp to manually reset their status.

- Moving the latest /var/log/secure file and restarting syslogd.  The
entry for authpriv.* remains in /etc/syslogd.conf and other syslog files
are still working file.

- Resetting the date manually (timezone is correct).

The only things that make be believe that this might be a security
attack is the fact that /var/log/secure no longer works AT ALL and that
I can't get a traditional list of current logins any more.  The other
things I could pass as a fluke or bug in the system.  But I have found
nothing else out of the ordinary.  The future file date issue is
terribly inconvenient, as well as the fact that I can't get a list of
current logins.

I found nothing on the Red Hat web site regarding these specific
issues.  I did find a new syslogd there, but the fixes don't seem to
address specificically the problems with no security logs.

Any ideas?  I'm at the edge of my wit...

Doug

2. how to change color depth in Redhat

3. Bizarre kernel problem

4. X vs PPPd

5. Bizarre shell problem. Please help

6. monitoring tool found

7. Bizarre ps problem

8. Number 9 GXE64pro

9. HELP! Bizarre FTP Problem!

10. Bizarre INFO problem with 3.1.5

11. bizarre rdist problems

12. Really bizarre semi-intermittent "cdrecord" problem

13. Really bizarre printing problem - RH 5.2