From owner-freebsd-stable Tue Mar 2 17:26: 8 1999 Delivered-To: freebsd-stable@freebsd.org Received: from mail.HiWAAY.net (fly.HiWAAY.net [208.147.154.56]) by hub.freebsd.org (Postfix) with ESMTP id 23AE214DF4 for ; Tue, 2 Mar 1999 17:26:05 -0800 (PST) (envelope-from dkelly@nospam.hiwaay.net) Received: from nospam.hiwaay.net (tnt8-216-180-15-176.dialup.HiWAAY.net [216.180.15.176]) by mail.HiWAAY.net (8.9.1a/8.9.0) with ESMTP id TAA21370; Tue, 2 Mar 1999 19:25:44 -0600 (CST) Received: from nospam.hiwaay.net (localhost [127.0.0.1]) by nospam.hiwaay.net (8.9.2/8.9.2) with ESMTP id TAA10091; Tue, 2 Mar 1999 19:25:19 -0600 (CST) (envelope-from dkelly@nospam.hiwaay.net) Message-Id: <199903030125.TAA10091@nospam.hiwaay.net> X-Mailer: exmh version 2.0.2 2/24/98 To: Tom Cc: freebsd-stable@FreeBSD.ORG From: David Kelly Subject: Re: rc5des slows tape thruput In-reply-to: Message from Tom of "Mon, 01 Mar 1999 19:59:32 PST." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 02 Mar 1999 19:25:19 -0600 Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Tom writes: > > The drive at home has 1M of internal RAM, the DDS-3 drives at work have > > 2M buffers. It would appear FreeBSD doesn't make use of these buffers. > > I don't see how that could be. SCSI tapes are SCSI tapes. Any > buffering done by the device should be invisible. There might be a SCSI > mode page flag to turn it on or off though. What if you write 1MB ahead into the tape drive, then hit the end of tape? How would one attempt to recover from that? It seems the Un*x architects intend to be able to recover from that situation altho in practice its unreliable. It seems only one block at a time is given to the drive, and the drive doesn't ask for more until the first is safely on tape. I have a Seagate/Conner/Archive DAT book that is begging me to read some more. > > A couple of percent. Using dd to read /dev/zero and write /dev/null > > yields 208M/sec w/o rc5des running, 203M/sec with, writing 100000 > > blocks of 10k (roughly 1G) so the test runs long enough to be valid. > > Those figures are pretty weird. 208MB/s or ? That is a pretty fast > tape drive. /dev/zero isn't a good test source due to drive compression > and encoding stuff. Read the above again. I was writing to /dev/null. The experiment suggests /dev/rsa0 is blocking on writing of only one block, while /dev/null doesn't block. rc5des hurt writing to /dev/null by 2.5% but hurts writing to /dev/rsa0 by 50%. The suggestion to hack HZ in the kernel isn't a bad idea at all. Other than the little problem the machine at work that I really want to apply this to is no longer under my control. Its owner moved it to another facility where I have to visit it as a common user. :-( -- David Kelly N4HHE, dkelly@nospam.hiwaay.net ===================================================================== The human mind ordinarily operates at only ten percent of its capacity -- the rest is overhead for the operating system. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message