From owner-freebsd-scsi Thu Mar 6 12:00:11 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id MAA07953 for freebsd-scsi-outgoing; Thu, 6 Mar 1997 12:00:11 -0800 (PST) Received: from Sisyphos.MI.Uni-Koeln.DE (Sisyphos.MI.Uni-Koeln.DE [134.95.212.10]) by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id LAA07841; Thu, 6 Mar 1997 11:59:45 -0800 (PST) Received: from x14.mi.uni-koeln.de (annexr2-41.slip.Uni-Koeln.DE) by Sisyphos.MI.Uni-Koeln.DE with SMTP id AA20411 (5.67b/IDA-1.5); Thu, 6 Mar 1997 20:59:39 +0100 Received: (from se@localhost) by x14.mi.uni-koeln.de (8.8.5/8.6.9) id UAA04017; Thu, 6 Mar 1997 20:58:56 +0100 (CET) Message-Id: <19970306205735.TD54148@x14.mi.uni-koeln.de> Date: Thu, 6 Mar 1997 20:57:35 +0100 From: se@freebsd.org (Stefan Esser) To: marck@gw.rinet.ru (Dmitry Morozovsky) Cc: freebsd-scsi@freebsd.org, freebsd-hardware@freebsd.org Subject: Re: NCR 510 troubles References: <199703051649.TAA29903@gw.rinet.ru> X-Mailer: Mutt 0.60_p2-3,5,8-9 Mime-Version: 1.0 In-Reply-To: <199703051649.TAA29903@gw.rinet.ru>; from Dmitry Morozovsky on Mar 5, 1997 19:49:33 +0300 Sender: owner-freebsd-scsi@freebsd.org X-Loop: FreeBSD.org Precedence: bulk On Mar 5, marck@gw.rinet.ru (Dmitry Morozovsky) wrote: > Hello there. > > We've got news server (2.1.5, partially upgraded to .7 -- at least, kernel > sources upgraded fully) with onboard ahc & additional ncr 510, which holds 3 > disks with a ccd on it. ncr configured for 2 tags on bootup. > > After upgrading newsserver from cnews to inn, > regularly, there are very sad things such as: > > Mar 5 16:38:50 gw2 /kernel: ncr0: restart (ncr dead ?). > Mar 5 16:38:51 gw2 /kernel: sd5(ncr0:4:0): UNIT ATTENTION asc:29,0 > Mar 5 16:38:51 gw2 /kernel: sd5(ncr0:4:0): Power on, reset, or bus device reset occurred > then system either crashes or one of ncr disks become missing. > (ahc0:0:0): "CONNER CFP1060S 1.05GB 2135" type 0 fixed SCSI 2 > (ahc0:1:0): "CONNER CFP2105S 2.14GB 2B4B" type 0 fixed SCSI 2 > (ahc0:2:0): "Quantum XP34300 81HB" type 0 fixed SCSI 2 > (ncr0:0:0): "CONNER CFP1060S 1.05GB 245F" type 0 fixed SCSI 2 > (ncr0:1:0): "CONNER CFP1060S 1.05GB 2035" type 0 fixed SCSI 2 > (ncr0:4:0): "QUANTUM EMPIRE_1080S 1101" type 0 fixed SCSI 2 > Any proposals? Hmmm, this is quite a number of disk drives ... Are they all mounted in the same case and connected to the same power supply ? Are you sure, your power supply is capable of delivering sufficient 12V current, if all drives are seeking at a time ? I guess that your seek rate increased when switching from c-news to inn, and since you are using ccd, there is a very high probability, that all drives are seeking at the same instance of time. This may lead to current spikes, which possibly overload your power supply ... I have ordered a new play-ground system for NCR driver tests and development, which hopefully will be delivered soon (finally). Improving the recovery from error situations is on top of my ToDo list, actually. But my current job does in part depend on the fact, that my (currently only) system is working reliably and without disruption, and as you can imagine, this prevents me from doing any driver tests :) BTW: You should upgrade from 2.1.5, for various reasons, and you'll find, that the NCR driver has become more reliable in case of a transient hardware failure, than it used to be. This is in part due to changes in the generic SCSI code, and I'm not sure, whether those made it into 2.1.7, but at least under 2.2, many more error situations should not be fatal. Regards, STefan