Arch Linux

steelneck

Member

Registered: 2011-08-26

Posts: 57

Strange boot fail.

About 7 times out of ten my machine boot up in some strange way, i  really do not understand what is going wrong. It started already at first boot after fresh install, but then i blamed it on some combination old lilo together with new kernel and some sort of bios-thing. But i have still troubble booting up Arch and there is no problem what so ever booting my old Slackware 12 with the same bootloader on the same disk, just another partition.

When things go wrong it boots blistering quick, really amazing quick. But i cannot log in as my regular user and when i log in as root i get the message "Non-standard uts for running kernel" and a code 196608, all my settings are gone and most installed programs wont start, it is like it is running from a ram disk, though my root partition is mounted but i have no path env. When the boot fails, all i have to do is to get into bios and then save and exit _not_ change anything! and then it boots fine. And when the boot go fine, i am _not_ greeted by that message upon login in as root.

There seem also to be some kind of time aspect in this. Now when i have got the machine up and running fine, i can reboot and all goes well, but if i turn it off for some hours, the trouble is back again.

And as i wrote, there are no problem with my old Slack 12 on the same disk and there was no problem with Slack 13 that Arch replaced. Yeasterday i tried to update the kernel to 3.0.4, but no difference. I really do not have a clue where to even start chasing this bug.

ewaller

Administrator

From: Pasadena, CA

Registered: 2009-07-13

Posts: 20,354

Re: Strange boot fail.

Anything hooked to your computer at boot?  USB drive, phone, mp3 player?  Any optical disks in the drive?

After boot (both "good" and "bad") what does uname -a report?

---

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

steelneck

Member

Registered: 2011-08-26

Posts: 57

Re: Strange boot fail.

Only a wacom tablet, but that has been there for some years not cusing any trouble earlier, not now either since that old Slack 12 boots just fine.

On a good boot, like now, uname -a says:
"3.0-ARCH #1 SMP PREEMPT Tue Aug 30 07:32:23 UTC 2011 i686 AMD Athlon(tm) 64 Processor 3000+ AuthenticAMD GNU/Linux"

But as i wrote, it may take some time to get a bad boot.

steelneck

Member

Registered: 2011-08-26

Posts: 57

Re: Strange boot fail.

Some new info.

Now when trying to get it to boot bad on purpose i discovered that the boot goes wrong at cold boot, that is when the machine has been powerless, if i just make a init 6 reboot from a good start it will reboot fine.

uname -a from a bad boot says the same as above.

I also discovered that the disks get mixed up. On cold (bad)boot sdb becomes sda and vs. on good boot. I do not get how this is possible since that old Slack 12 with its root on sda2 boots just fine in either case, no mix up of disks (using the same grub). To get Arch booting and have the right disk as root i have to go into bios and just press F10 (save and exit), not change anything, then sda is sda and not sdb.

Last edited by steelneck (2011-09-03 23:35:16)

falconindy

Developer

From: New York, USA

Registered: 2009-10-22

Posts: 4,111

Website

Re: Strange boot fail.

In other words, you're not identifying the root by UUID and relying on naming which is assigned through an asynchronous process (udev/devtmpfs).

---

blog | ±github

steelneck

Member

Registered: 2011-08-26

Posts: 57

Re: Strange boot fail.

Identify root by UUID? Where and to what? fstab? Remember, my old OS boot just fine from the same disk with the same bootloader. Is this some new req. of the 3.0 kernel or is it something else?

ewaller

Administrator

From: Pasadena, CA

Registered: 2009-07-13

Posts: 20,354

Re: Strange boot fail.

https://wiki.archlinux.org/index.php/Fstab#UUID

And then update /boot/grub/menu.lst to refer to the disks by these UUIDs

Last edited by ewaller (2011-09-04 01:19:55)

---

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

steelneck

Member

Registered: 2011-08-26

Posts: 57

Re: Strange boot fail.

I will try, but i am a bit skeptic since i cannot see how this explains the "save and exit without change in bios", especially not in grub since sda is sda when i chose to boot Slackware on sda2 instead of Arch on sda1. Also, if i simply disconnect the power from the sdb disk, the problem is gone

ewaller

Administrator

From: Pasadena, CA

Registered: 2009-07-13

Posts: 20,354

Re: Strange boot fail.

I suspect it has to do with a drive not being ready in time at POST.  In its absence, BIOS assigns the first spot to the drive that is ready.  Spending time in the BIOS setup screen allows the slow drive to be ready when the order is assigned.  Just a hypotheses.

---

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

steelneck

Member

Registered: 2011-08-26

Posts: 57

Re: Strange boot fail.

I have just changed to UUIDs in fstab only, not in grub, and it seem to have done the trick. At sometime over the yers i guess i picked up to only change one thing at the time when trying to single out a problem.. Now i have done both a cold and a "warm" reboot and it seem to work OK. Thanks for the tip, though i still wonder exactly what is causing the mixup in Arch but not in my old Slackware.