Server does not boot anymore

English Install/Boot/Login
1

Hi,

I have a big problem.
My server with Opensuse 11.1 doesn’t boot anymore.

This afternoon I’ve installed imagemagick over yast. After that the server still works for a couple of minutes. Then it stops working (I’m not sure if it’s because of the installation). so I had to reboot it.
But it never came up again.
It doesn’t boot anymore.

I get the boot menu and can make my choice. But it doesn’t matter what I choose (standard or failsagfe) the only message I get is:

  Booting 'Failsafe -- openSUSE 11.1 - 2.6.27.48-0.2'

root (hd0,0)
 Filesystem type is ext2fs, partition type 0xfd
kernel /vmlinuz-2.6.27.48-0.2-default root=/dev/md1 showopts ide=nodma apm=off
noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 x11
failsafe
   [Linux-bzImage, setup=0x3000, size=0x269760]
initrd /initrd-2.6.27.48-0.2-default
   [Linux-initrd @ 0x37c9f000, 0x350b4a bytes]

After initrd the startup ends. Even if I change splash to verbose in menu.lst.

My hoster gives me access about a rescue-system independent from my installation and I have access to my harddisks, configs and logs.

My boot.log is completly empty and the only msg in boot.msg from today is::


Boot logging started on /dev/tty1(/dev/console) at Wed Oct  6 16:37:56 2010

Master Resource Control: previous runlevel: 3, switching to runlevel: 6
Shutting down Qmail <notice>killproc: kill(5154,15)
<notice>killproc: kill(5154,9)
done
Shutting down CRON daemon<notice>killproc: kill(4780,15)
done
Shutting down mailman<notice>killproc: kill(5252,15)
<notice>killproc: kill(5252,9)

But this is hours ago. And I tried to reboot the server a couple of times. These tries are not entered in any log.

My websites are down too. So nothing seems to work.

Any ideas?

Thanks

Stefan

2

Possibly a corrupted kernel or initrd. Probably nothing to do with anything you did. Are you able to check the harddisk for errors?

3

So, really, you are saying the computer boots, but the kernel no longer loads. Correct? First, do you have any extra kernels to try in your menu.lst file? Second, do you have your original openSUSE 11.1 boot disk? Why not try a reinstall, but doing an upgrade, that will reload the original kernel. Here is an an example by caf4926 of doing this in openSUSE 11.3:

Picasa Web Albums - carl fletcher - DVD Kernel Fix

Thank You,

4

I’ll do a badblocks-check. But this will take a couple of time. I will post as soon as I have results. Thanks

I have no access to a boot disk. The server is hosted. The only access I have is a rescue system (ubuntu) which gives me access to my harddisks. I’ve tried chroot on my harddisk and start yast, but this didn’t work.
The other option is to reinstall my SUSE 11.3 from a standard image. But this means a loss of data (about 16 hours from my last backup - if the backups works correctly)

Thanks.

5

Maybe you should install a good kernel and initrd anyway and boot that.

6

Hello Stefan, i read:

My hoster gives me access about a rescue-system independent from my installation and I have access to my harddisks, configs and logs.

So this is a remote system? Would be good to know what kind of access you still have (or have not). Are backups a problem (have it: yes/no)? I think the breakdown is not caused by the installation of ImageMagick, but a pure coincidence.

Is this a virtual server? I ask, because the startup message you see would mean that - on a physical server - grub is working, and that’s quite a lot (this would exclude some HW failures).

[Ooops, I am way in the past. Most questions have already been answered; sorry for the noise]

7

I don’t know how to do that manually. mI’ve only done this via yast/you. I never compiled a kernel by myself.

Some additional information:
The server is a dedicated server. My only access is via remote console through putty.

I’ve made backups using the plesk backup. I know this only works for my websites and plesk things, but these are the most important.
But I’m afraid, that there could be an error restoring the backup (this is always possible). I mean it should work, but what if not? I’ve made tar -cvzf on my root and put this file on a backup space. Is this enough to restore my websites and databases even if pleskbacup doesn’t work.

I’m thinking about restoring a standardimage, making updates via YOU and then restore my beackup.
This means a loss of data, but I think it’s less worth than not beeing reachable for days (even via email).

Is the tar enough for failsafe?
The harddisks seemto be OK.

Stefan

8

It’s a bit tricky, but you can extract a kernel RPM (say the original 11.1 one) into the server’s disk using rpm2cpio, then edit /boot/grub/menu.lst to add that to the list, or just boot it from the interactive GRUB menu (I assume you have some kind of virtual console access).

Something like this:

cd /
rpm2cpio kernel-default-blah-blah.rpm | cpio -idmv
9

After a couple of additional hours of trying to solve the problem, I decided to reinstall the standardimage, do the updates and restore my backup.

OK - there was a loss of data. But otherwise I was not reachable for days (not my websites or myself by email).

Thanks to all, who gave me hints to solve my problem.

Stefan