Xref: utzoo comp.unix.admin:1561 comp.unix.questions:30269
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!csn!ncar!gatech!prism!eecom.gatech.edu!flur
From: flur@eecom.gatech.edu (Peter W. Flur)
Newsgroups: comp.unix.admin,comp.unix.questions
Subject: Performance of Dump
Message-ID: <26030@hydra.gatech.EDU>
Date: 11 Apr 91 12:21:07 GMT
Sender: news@prism.gatech.EDU
Followup-To: comp.unix.admin
Organization: Computer Engineering Group, Georgia Tech, Atlanta, GA 30332-0250
Lines: 28

I am using an exabyte attached to a Sun for dumps of several different types of machines
on our network including Suns, NeXT's, Dec's, and HP workstations.  When I dump from a 
Sun to the dump host (a Sun), every thing works great and the transfer  is blazingly
quick.  However, when I do a dump from a NeXT (running 1.0a)  or one of the HP workstations
(running HP/UX), the dump works, but it is noticeably slower than dumping another Sun
with larger disks.   For example, a full dump of a NeXT with a 330MB (280MB used) drive 
takes roughly (these are estimates but close in magnitude) 45-60 minutes.  A Sun with a 
660MB drive (270MB used) takes less than 15 minutes.  Both machines are dumped from the
tape host through a shell script as follows:
rsh workstation rdump 0nubsdf 1024 6000 54000 tapehost:/dev/exabyte /partition
where the obvious substitutions are made with the appropriate names.

Can anyone offer a reason why these other machines are significantly slower than the
suns, and more importantly, can someone offer a solution to speed them up?

Thanks,
Peter

-- 
-----------------------------------------------------------------------
		    Peter Flur, Research Engineer
		   Georgia Institute of Technology
      School of Electrical Engineering, Atlanta, GA  30332-0250
		     E-MAIL: flur@ee.gatech.edu
			PHONE: (404) 853-9355
-----------------------------------------------------------------------
			   DUKE UNIVERSITY
		  1991 NCAA BASKETBALL CHAMPIONS!!!