[SOLVED] No init found. Try passing init= bootarg

Hello,
I've been running mint9 xfce as my main distro for ages and it's proved bombproof until about an hour ago.
Everything locked and eventually I had to hit the power button. On attempting a re-boot I get the usual pretty green Mint logo and then a black screen with the above message (No init found etc.)no mention of kernel panic.
Recovery mode leads to the same result and using a live distro (puppy) results in an "unable to mount" message.
Fortunately everything is backed up to another hdd up until 2 days ago but there are several emails (this is a business PC) which I don't want to lose by a re-install. Can anyone offer advice?

Have you tried doing an fsck on the root filesystem? You'll need to boot the rescue medium to do this of course.

This does seem to be a corrupted filesystem problem. The result of "fdisk -lu /dev/sda" (replace sda with the proper drive if different) should be posted as a start, so we can see what partitions you have, or if the partition table is corrupt. Then for each partition not listed as swap, try to mount it in read-only mode to get error messages. Just mount on /tmp for now like "mount -r /dev/sda1 /tmp" and show what messages are producted.

Later you can try to do the appropriate "fsck" command to attempt to recover the filesystem. Don't be in a hurry to do that, since many kinds of corruption could actually get worse by trying that.

What I usually do in a case like this is remove the drive and put in a replacement (same size or larger). I then install a new system to the new drive. After that, I attach the old drive to the connections for a 2nd or 3rd drive as the case may be. Then I have a full system with its own permanent space to save stuff I can recover. I can defer recovery to later if getting the system back up is the urgency.

Hello and thanks,

I think the last line maybe a clue but I don't know where to go next (I've used Linux for years and I'm still useless)

Disk /dev/sda1: 489.3 GB
255 heads, 63 sectors/track, 59482 cylinders, total 955592704 sectors
Units = sectors of 1 * 512 bytes = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000

Disk /dev/sda1 doesn't contain a valid partition table

Originally Posted by starkadder Hello and thanks, I think the last line maybe a clue but I don't know where to go next (I've used Linux for years and I'm still useless) Disk /dev/sda1: 489.3 GB 255 heads, 63 sectors/track, 59482 cylinders, total 955592704 sectors Units = sectors of 1 * 512 bytes = 512 bytes Sector size (logical/physical): 512 bytes / 512 bytes I/O size (minimum/optimal): 512 bytes / 512 bytes Disk identifier: 0x00000000 Disk /dev/sda1 doesn't contain a valid partition table

That's a partition you did it on, not the whole disk. Do the command on the whole disk by not having the '1' in the device name. E.g. do this:

fdisk -lu /dev/sda

Yes, thank you Skaperen,
This is the result:

knoppix@Microknoppix:~$ fdisk -lu /dev/sda

Disk /dev/sda: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders, total 976773168 sectors
Units = sectors of 1 * 512 = 512 bytes
Disk identifier: 0x0003f771

Device Boot Start End Blocks Id System
/dev/sda1 * 2048 955594751 477796352 83 Linux
/dev/sda2 955596798 976773119 10588161 5 Extended
/dev/sda5 955596800 976773119 10588160 82 Linux swap / Solaris
knoppix@Microknoppix:~$

Most live distros can't open the drive and puppy won't let me copy the terminal, good old Knopix.
Scratching around with a recovery programme produced the following:

Sector read Error at CHS=(29519;95;20;LBA=474228739)


All input gratefully acknowledged.

Hi,
I've managed to retrieve my home folder intact by using the script below.

titus@bisto ~/Desktop $ sudo fsck /dev/sdb1
fsck from util-linux-ng 2.17.2
e2fsck 1.41.11 (14-Mar-2010)
/dev/sdb1: recovering journal
Error reading block 59278337 (Attempt to read block from filesystem resulted in short read). Ignore error<y>? yes

Force rewrite<y>? yes

/dev/sdb1: clean, 171608/29868032 files, 16250925/119449088 blocks
titus@bisto ~/Desktop $

The rest of the installation appears undamaged but it won't boot. I could do a clean re-install but it would be nice if I could repair the existing set up.
Any ideas?

Originally Posted by starkadder The rest of the installation appears undamaged but it won't boot. I could do a clean re-install but it would be nice if I could repair the existing set up. Any ideas?

What happens when you try to boot?

Hello eeekster,
Thanks for the response.
Basically, nothing happens, the PC posts, beeps and then it's a cursor on a black screen. I've shifted the Boot folder from a working installation of the identical distro but there is no change. It's sort of academic really as I have retrieved the files I need and I'm typing this on an install of Mint 9 xfce on another drive, I just don't like mysteries and although I'm old nobody is too old to learn.
As ever, any input is appreciated.

You will need another disk (or two) to do recovery by copying what you can get out of the bad disk. I'd copy the raw sectors, first, working around the bad one as much as possible. A copy that can be accessed without an I/O error might work easier than one the hardware interferes with. If that mounts, the copy the files you can get out of it.

Simple plan is to get a new disk to put a new OS on, and have it also be the place to do the file copy of old files to.

Do not write to the bad disk in any way. Do mounting in read/only mode.

Hello Skaperen,
Thanks once again.
I've recovered the home folder intact and as this contains all the files and emails etc. that were not backed up to the other disk I've done a clean install (on a fresh hard drive), transferred the entire home folder plus a few visual tweaks and I'm running that.
I tried "grub-install" on the un-bootable drive and got the message that all was well but it still wouldn't play. The Bios recognises the drive and from the new "main" drive I can access it without problem now, why it won't boot is a mystery to me but it's not important. I'll probably format the original drive and start again.
Thank you for your help, and thanks too to the others who showed an interest in my problem.
It's refreshing to use a forum where the experts don't treat the non-experts like idiots.
Cheers! Titus.

Originally Posted by starkadder I tried "grub-install" on the un-bootable drive and got the message that all was well but it still wouldn't play. The Bios recognises the drive and from the new "main" drive I can access it without problem now, why it won't boot is a mystery to me but it's not important. I'll probably format the original drive and start again. Thank you for your help, and thanks too to the others who showed an interest in my problem. It's refreshing to use a forum where the experts don't treat the non-experts like idiots. Cheers! Titus.

The "grub-install" script and/or the underlying tools it uses, may not check every possible failure mode. More likely, they check for just enough to see if all the previous steps have been done, or not. That would be steps like the drive exists, has a partition table, has the target partition, and the grub package has been installed. That would work fine in even a manual install scenario. But it could easily fail in a scenario where some little obscure thing is broken (even though everything else that would installed at that time is OK ... e.g. it was fooled because it is not practice to test every little detail).

Wanting to know why it fails is one thing and I can appreciate the learning aspect of that. OTOH, a lot of use geeks will just accept that some little thing is broken somewhere, and focus on making it work no matter what is broken, if possible.

What I'd end up doing is hard-zeroing the drive, or at least the beginning part of it. That would wipe out the partition table and the MBR boot sector (which may or may not be the broken detail in your drive). It could also wipe out one or more partition, depending how exhaustive I want to be. I usually do wipe the whole drive unless there is data I want to keep.

Then I'd just install again. Since it is on a disk that is wiped to binary zero, then any scripts or programs doing installs could not assume some part is installed and skip it. It would be forced to definitely install everything that is needed.

At this point is where we part ways. I gave up on grub and have moved to syslinux, at least for the interim. I'm hoping to end up writing my own bootloader if I get the time to do so. In the mean time, syslinux is doing well for me. I've yet to switch my Mint system over to it, but I have done so with Slackware (it uses lilo by default but includes syslinux) and Ubuntu. I will be using syslinux for my bootable memory stick builder project.