I made up my own "crashme" based on something posted earlier, but it's substantially the same. I left it running for the weekend - I'll on Monday night how it worked out. When I left work this morning it had completed 12 cycles, copying about 6Gb of distfiles to an lvm partition, then doing the tar, &c there. Each loop takes about 25 minutes.
What I *don't* have is any kind of odd external hardware, just the internal disks (10 * 18Gb). I think I saw someone refer to an external array - could that be part of the trigger?? I have no clue, as I'm not much of a kernel hacker. I was just thinking that *any* input might be helpful in tracking down the 2.6.x instability issues, and my E450 isn't doing much other than tracking portage updates at the moment.
I also have a couple of non-smp Ultra1+Creator3D's that seem to be happy running 2.6.x. Again, no odd external hardware, just whatever came in the original box.
As I said before, I'm happy to risk destablizing a system if it'll help track down the stability issues. It's about time I started giving *something* back...:)
Ralph
-----Original Message-----
From: Josh Grebe [mailto:squash@gentoo.org]
Sent: Sat 7/9/2005 8:25 AM
To: gentoo-sparc@lists.gentoo.org
Subject: Re: [gentoo-sparc] Kernel Panic Problems
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi Ralph,
I also have an E450, and had a couple hundred days uptime...
but the bug eventually got me. Compared to my test box,
which only goes for 20 minutes to an hour or so, the E450
was suprisingly stable. It may have to do with the CPu
speed, as my E450 has 4x400 and my test box has 4x440....
but at this point its had to say.
Meanwhile, my 450 is running 2.6, becuase the bug was very
hard to trigger.
As best as I can tell, any SMP US2 box is affected.
If you want to use the same crashme.sh that we are using,
http://dev.gentoo.org/~squash/crashme.sh although it sounds like you may already have it.
Josh
Mitchell, Ralph wrote:
> I haven't really been following this thread closely, but I was wondering - is this a problem with 2.6.x on any kind of UltraSparc, or is there a definite set of hardware that has issues??
>
> Maybe I've just been insanely lucky, but I had 2.6.10-gentoo-r6 running on an E450 (4 x 400MHz, UltraSparcII (Blackbird)) for over 150 days uptime. In that time I've rebuilt the toolchain a couple of times, as well as Firefox, kdelibs & kdebase, and xorg-x11 and a bunch of miscellaneous stuff. Right now I have 6 days uptime on 2.6.12-gentoo-r3, and it's been rebuilding xorg-x11 for a couple of hours.
>
> If I *should* be having problems, but am not for some odd reason, maybe folks can suggest stuff for me to try out?? I'm willing to try to break it, if that would help resolve the problems other people are experiencing.
>
> I'm just about to start up a 'crashme' script, doing the copy/tar/untar/delete loop thing, just to see what happens.
>
> Ralph Mitchell
>
>
> -----Original Message-----
> From: Paul Heinlein [mailto:heinlein@madboa.com]
> Sent: Fri 7/8/2005 5:24 PM
> To: gentoo-sparc@lists.gentoo.org
> Subject: Re: [gentoo-sparc] Kernel Panic Problems
>
> On 7/8/05 2:26 PM, Nick Smith wrote:
>
>>I just installed gentoo stage1 on an Ultra10, with kernel
>>2.6.12-gentoo-r4, it starts up fine, but then after a couple minutes I get:
>>
>>Kernel Panic - not syncing: Aiee, killing interrupt handler!
>>
>>And the system locks.
>
>
> Me, I'd install a 2.4 kernel. You'll find some recent threads (e.g.,
> "2.6 kernel and docs") on this list that deal with the troubles people
> have with 2.6.
>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird -
http://enigmail.mozdev.org iD8DBQFCz9BeFAhB33r2ACYRAiYEAJ0Q6i7sTEXGVCpjCtYIdmry2+YKrwCeIrci
VTrgODPb9+4fsKdvGo32/i4=
=H3t/
-----END PGP SIGNATURE-----
--
gentoo-sparc@gentoo.org mailing list