From owner-freebsd-scsi Mon Aug 17 16:11:56 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA15107 for freebsd-scsi-outgoing; Mon, 17 Aug 1998 16:11:56 -0700 (PDT) (envelope-from owner-freebsd-scsi@FreeBSD.ORG) Received: from drew.doit.wisc.edu (drew.doit.wisc.edu [144.92.30.77]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id QAA15093; Mon, 17 Aug 1998 16:11:54 -0700 (PDT) (envelope-from jeffb@doit.wisc.edu) Received: (from jeffb@localhost) by drew.doit.wisc.edu (8.8.7/8.8.7/drew-v2) id RAA02163; Mon, 17 Aug 1998 17:49:14 -0500 (CDT) (envelope-from jeffb) From: Jeff Bartig Message-Id: <199808172249.RAA02163@drew.doit.wisc.edu> Subject: Re: tape device busy In-Reply-To: from Chris at "Aug 17, 1998 5:20:45 pm" To: chris@exit109.com (Chris) Date: Mon, 17 Aug 1998 17:49:14 -0500 (CDT) Cc: freebsd-isp@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL45 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Chris said on freebsd-isp: > every once in a while, i get this error when i go to perform a backup: > > mt: /dev/rst0: Device busy > > Is there anyway to reset the device on the machine without rebooting it? I'm copying this response to freebsd-scsi also, since that is where I thought to search when I was having this problem recently. I have been having possibily the same problem with 3 Seagate STD28000N DAT drives. Two drives had version 6500 firmware, while the other had version 6280. Processes like mt and rmt (rdump or gtar initiated) were hanging on three similarily configured FreeBSD 2.2.6-RELEASE systems. Once the process hung, I was unable to kill it. Running ps showed that the process was in the "D" STATe (disk or other short term, uninterruptible wait). Since these processes still had the /dev/rst0 device open, I would get the "Device busy" errors. The only solution I could find was to reboot the system. There wasn't any pattern to when the problem would occur. Some backups would work fine and the problem wouldn't appear for days at a time. Searching with DejaNews came up with some postings suggesting that the Archive Python DAT drives (which the Seagate is a decendant of) had problems with command tag queueing. I sent an e-mail to Seagate tech support about the problem. Seagate didn't answer many of the questions I asked, but did include the firmware upgrade I requested. I installed the firmware upgrade to version 658A. It required using a DOS boot disk with ASPI drivers loaded on it. I have now been running for 8 days using the new firmware. I have not had a single problem with the mt or rmt processes locking up. My busiest tape drive writes multiple tar/dumps onto 3 or more tapes a night. Since I did the upgrade, I seem to be having many less tape problems overall. I'm going to run for a while longer before I assume that I have found the solution to my problems. Hopefully this is the answer. Jeff -- Jeff Bartig | University of Wisconsin - Madison 1210 W Dayton, Rm 3219 | Division of Information Technology Work Phone: (608) 262-8336 | Network Engineering Technology E-Mail: jeffb@doit.wisc.edu | To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message