Issues with kernel loading past IO Scheduler

Insanity5902

I am going to do the best I can to explain this, but I am not sure how well I will do at it as I don't fully understand the problem.

I guess I will try to explain the issue first and then give you some files.

First off, the gentoo livecd 2005.1 works without a problem, boots fine every time. The issue I am having is when grub loads my kernel. It freezes a lot right after it loads the IO Scheduler and after looking it is right before it loads the RTC timer. I have notice that off of a fresh boot it will stop right there everytime, and then when I reboot it goes through fine. Now if I reboot again, it will hit that point every single time there after. I have to turn the machine off, what a few seconds, turn it on, let it freeze reboot and then it will load, and it will go through that cycle everytime.

I have tried the different IO Schedulers and RTC timers. The only thing I haven't tried is leaving everything for RTC and IO schedulers out. (I will try that here in a little bit).

Another thing to note, is everytime it hits that spot in when the kernel loads, whether it hangs there or not, is it will beep 3 times.

So here are the files that might be of intereset

dmesg
kernel config
lshw

Even though I included lshw, I thought I would list a few of the hardware for the forum searches
2 x Intel PIII Xeon
Intel SCSI Raid
Intel SCB2 Motherboard

If you need anything else let me know
_________________
Join the adopt an unanswered post initiative today

Insanity5902 · Posted: Wed Feb 08, 2006 4:21 pm Post subject:

After talking to the guy who owned it before me, he said when he ran ESX on it, they told him he had to disable APIC to get it to work.

Well I dove into the kernel Docs to read about it, and with SMP machines, APIC is enabled by default, and the options doesn't show up in the menuconfig when you choose SMP. I could go into the .config and disable it by hand, but I only have ssh access at the moment and don't want it to crash while I am out.

Has anybody had any issues with APIC on SMP machines before?
_________________
Join the adopt an unanswered post initiative today

Insanity5902 · Posted: Thu Feb 09, 2006 1:29 pm Post subject:

Well it seems if you have SMP enabled, whether or not the .config has APIC enabled or not, it throws it in by deafult. I am going to try compiling a kernel without SMP (and therefore without APIC) and see how that goes.
_________________
Join the adopt an unanswered post initiative today

bollucks · l33t Joined: 27 Oct 2004 Posts: 606

goofygrin · n00b Joined: 07 Feb 2006 Posts: 11

I ran into the same problem with a single core/single cpu machine after swapping the motherboard (everything else the same).

I didn't have time to futz with it, so I just swapped the motherboard out to the old one and said forget it.

I tried no smp, no acpi, etc. to no avail.

I'll keep an eye on this thread though to see what you get.

Insanity5902 · Posted: Thu Feb 09, 2006 7:24 pm Post subject:

Insanity5902 · Posted: Fri Mar 17, 2006 9:04 pm Post subject:

Well, a month later and I finally got back to screwing with it. I was going to try a different raid card, but the motherboard would never post with it in, so I went back to the Intel card and decided to do some kernel testing, so I loaded up the 2.6.16-rc6 vanilla-sources and sure enough, everything is working fine.

I have no idea what it was, the only thing I really did differently was put some scsi debugging in the kernel. Other then that, it is pretty much exactly the same.

I am going to take this .config, and try it with the older sources to see how goes it, i even tried the 2.6.15-3 vanilla-sources with the same problem.

Now I am going to backtrack everything with the kernel config that works and see if it was kernel related or just my config related. I have had lots of problems with the 2.6.15 kernel on other machines, just weird issues, so I wouldn't be suprised to see this as the problem.

I will be sure to update this thread with my findings.
_________________
Join the adopt an unanswered post initiative today

dsd · Posted: Fri Mar 17, 2006 10:11 pm Post subject:

your problem is fixed in the more recent revisions of gentoo-sources-2.6.15
in future, please file bug reports at https://bugs.gentoo.org
_________________
http://dev.gentoo.org/~dsd

Insanity5902 · Posted: Sun Mar 19, 2006 3:08 pm Post subject:

I never filed a bug b/c I was never 100% sure what the problem was, at the time I didn't know if it was a kernel issue, or just a bad scsi card as I have never used that card before.

Next time I will go to the bug reports anyway. Thanks.
_________________
Join the adopt an unanswered post initiative today

tightcode · Tux's lil' helper Joined: 23 Mar 2004 Posts: 110

For those finding this later on and being dissapointed by the lack of specifics, here is some more detailed info.
- There doesn't appear to be any bug reports matching this problem (searching for io scheduler).
- It appears that starting from 2.6.15-r7 which is masked at the time of writing, the problem is solved (to give an actual revision to vague "the more recent revisions").

Cheers,

TightCode

dsd · Posted: Tue Mar 28, 2006 10:17 pm Post subject:

here is the bug report:
https://bugs.gentoo.org/show_bug.cgi?id=122277
there was another one too, but i cant find it.

-r7 is the latest 2.6.15 and therefore fixes all known instances of this issue. the actual fixes were spread over 2 or 3 revisions though.
_________________
http://dev.gentoo.org/~dsd