From owner-freebsd-questions Tue Jan 13 21:14:44 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id VAA17562 for questions-outgoing; Tue, 13 Jan 1998 21:14:44 -0800 (PST) (envelope-from owner-freebsd-questions@FreeBSD.ORG) Received: from Kitten.mcs.com (Kitten.mcs.com [192.160.127.90]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id VAA17555 for ; Tue, 13 Jan 1998 21:14:28 -0800 (PST) (envelope-from font@Jupiter.Mcs.Net) Received: from Jupiter.Mcs.Net (font@Jupiter.mcs.net [192.160.127.88]) by Kitten.mcs.com (8.8.7/8.8.2) with ESMTP id XAA18534; Tue, 13 Jan 1998 23:14:26 -0600 (CST) Received: from localhost (font@localhost) by Jupiter.Mcs.Net (8.8.7/8.8.2) with SMTP id XAA15798; Tue, 13 Jan 1998 23:14:25 -0600 (CST) Date: Tue, 13 Jan 1998 23:14:25 -0600 (CST) From: Font To: Doug White cc: questions@FreeBSD.ORG Subject: Re: How best to speed up rdumps? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk I did do some trials from a P100 IDE box to a PPro200 AHA2940UW box, both since upgraded a bit to 2.2.5R, and on the same small test partition got: -b 10 ~50K/s -b 16 ~50K/s -b 24 ~80K/s -b 32 ~220K/s -b 48 ~220K/s -b 60 ~220K/s The command used was dump 0au -b # -f remotehost:/dev/nrst0 /mountpoint So now I'm using -b 32 and I've seen dump rates of up to 400K/s on other dumps, which is acceptable. Thanks for the tips, they helped encourage me to experiment and learn. :-) dw A bug in my MUA causes news.announce.newusers font to be sent to beneficiaries and senders of UCE/SPAM. @ mcs.net Wishes are like dishes. On Mon, 12 Jan 1998, Doug White wrote: > Date: Mon, 12 Jan 1998 19:15:48 -0800 (PST) > From: Doug White > Reply-To: Doug White > To: Font > Cc: questions@FreeBSD.ORG > Subject: Re: How best to speed up rdumps? > > On Tue, 6 Jan 1998, Font wrote: > > > I'm dumping across a network to an Exabyte 8505, and while backups on the > > FreeBSD 2.2.2+ machine seem to go at 300-600Kbyte/s typically, backups > > across the network (from a 2.2.1R machine, so far) seem to only be getting > > around 50Kbyte/s. The network is 10baseT with a smattering of 100baseTX, > > and is fairly calm at the time of backup. The command I'm using is > > > > dump 0auf tapeserverbox:/dev/nrst0 filesys > > > > Should I be specifying a blocksize when doing backups over the network? > > If so, how would I go about selecting an ideal blocksize? Would I need to > > do several test dumps to determine this? Or should I be looking into > > something else for the bottleneck? > > I bet the tape isn't streaming. In that case, yes, you want to set a > blocksize. Try 10 to start with and work from there until you can get the > tape drive to stream (that is, not reposition the heads during the > backup). > > We had to use 10 to back up a 486/25 to a Pent/120 box with a Connor 4GB > drive. The default blocksize would shoeshine so much it would have taken > six months to back up 800 megs :-) > > Doug White | University of Oregon > Internet: dwhite@resnet.uoregon.edu | Residence Networking Assistant > http://gladstone.uoregon.edu/~dwhite | Computer Science Major > >