From owner-freebsd-alpha Wed Dec 8 13:18:26 1999 Delivered-To: freebsd-alpha@freebsd.org Received: from mailtoaster1.pipeline.ch (mailtoaster1.pipeline.ch [62.48.0.70]) by hub.freebsd.org (Postfix) with SMTP id 4D9D11504A for ; Wed, 8 Dec 1999 13:18:06 -0800 (PST) (envelope-from oppermann@pipeline.ch) Received: (qmail 59872 invoked from network); 8 Dec 1999 21:16:45 -0000 Received: from unknown (HELO pipeline.ch) ([195.134.128.41]) (envelope-sender ) by mailtoaster1.pipeline.ch (qmail-ldap-1.03) with SMTP for ; 8 Dec 1999 21:16:45 -0000 Message-ID: <384ECADC.D94CA286@pipeline.ch> Date: Wed, 08 Dec 1999 22:17:16 +0100 From: Andre Oppermann X-Mailer: Mozilla 4.03 [en] (WinNT; U) MIME-Version: 1.0 To: bjp aka rbw Cc: Mike Smith , freebsd-alpha@FreeBSD.ORG Subject: Re: 3.3-S crashing with "ncr0: queue empty." after ~2 weeks References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-alpha@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org bjp aka rbw wrote: > > On Mon, 6 Dec 1999, Mike Smith wrote: > > > > the next time, i decided to run Crack v5.0 to see if i had any weak > > > passwords on my system. this seemed to be semi-disk I/O intensive. it > > > crashed with the repeating "ncr0: queue empty." error after about 6 days. > > > > I've only seen this error in conjunction with failing NCR controllers; > > usually dying due to overheating in Multias. > > the box doesn't seem to be running much higher than room temp, so i kind > of doubt it's heat. (it's nothing like the multias... that's for sure.) > if it is heat, do you think that maybe the NCR controller is permanently > damaged? or do you think i should just look into cooling this thing? > > also, do you think there is a chance that this is an odd problem in > 3.3-STABLE? (from what i can gather, most of the poeple on this list are > running 4.0-CURRENT.) i will try 4.0-CURRENT when i can find the time and > see if that fixes it. > > oh yeah... and anyone have some sort of test i can run that is highly disk > intensive so i can try to force this problem to pop up? I had the same problem some month ago for some time, but I'm running -current on that box. At that time Stefan Esser did some modifications to ncr.c, I checked out the version before these changes but it didn't fix the problem. I suspect some other circumstances in the CAM stack to be the cause for it. "queue empty" is the default in a case statement. If someone can/want to track this down I'll help where I can. For some time now I havent got that error anymore, probably because I don't hit the HDD that hard anymore. -- Andre To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message