SUMMARY: Problems with Dump

From: Rubens Q. Almeida (queiroz@com.unicamp.ansp.br)
Date: Fri Mar 13 1992 - 05:53:43 CST


These are the replies I received concerning my problems with the dump program:
 
======================================================
Here are the options we use here:
 
   dump 0custdf 590 18 1200 /dev/rst0 /usr
 
The options are:
 
   0 - dump level (full dump)
   c - cartridge (sets some other options)
   u - update the "/etc/dumpdates" file
   s - size of the tape = 590 feet (really 600, but allow for junk)
   t - number of tracks = 18
   d - density = 1200 ftpi (flux transitions per inch)
   f - tape device file name = "/dev/nrst0"
 
Hope that helps.
 
   Drew
 
 +----+ +----+ +------------------------------+----------------------------+
 | \GO/ | | Drew Montag | djm@bioimage.millipore.com |
 ++ + \/ + ++ | Sr. Software Design Engineer | uunet!millidc!bioimage!djm |
  | |\ /| | | Bio Image Products | (800) 872-0071 (Toll Free) |
  | | \ / | | | 777 East Eisenhower Parkway | (313) 930-9900 (Voice) |
  | | \/ | | | Suite 950 | (313) 930-0990 (FAX) |
 ++ ++ ++ ++ | Ann Arbor, Michigan +----------------------------+
 | |BLUE| | | U.S.A. 48108-3258 | Are we having fun yet? |
 +----+ +----+ +------------------------------+----------------------------+
 
======================================================
 
From: mp@allegra.att.com (Mark Plotnick)
 
One version that's said to work is
        dump 0bcstf 126 600 18 /dev/rst0 filesystem
 
what's 19200? the density of your cartridges?
 
======================================================
 
I've been getting along for some time with dumps to 150MB
quarter-inch cartridges by specifying the c option, the t option
(tracks I think) with a value of 18, and a density of 1250. I got
these parameters from the net when QIC 150 drives were relatively
a new thing. I usually specify the length as 590 rather than 600
out of excess caution. Hope this helps.
 
Don Hooper, hoop@boulder.colorado.edu ...!uunet!ncar!boulder!hoop
 
======================================================
 
You need to say what version of the OS you are running to get the "optimal"
answer but in general you want to use a large blocksize parameter to speed
up the dump. Under 4.1.1, what I tend to use is
        dump 0cusdtf 590 1200 18 /dev/rst0 ....
The QIC-150 driver writes out 512k blocks independent of the blocksize and the
dump parameters. The other parameters are used by dump to estimate the
length of the tape so you want to choose parameters that dump will compute
to not actually exceed the available tape but come as close to the available
length as you can allowing for interrecord gaps, some bad spots on the tape as
well as dump header info.
 
One way to test your parameter choice is to type your favorite set of
parameters but to use /dev/null as the dump device (omit the u option so your
dumpdates file is not screwed up). Dump will output the number of megabytes
dumped and how many tapes it thinks it will need for that number of meg. Thus
you can see how many megabytes the parameter choices will yield. For the
parameters above, this trick gave me 331.71 meg on 2.34 tapes, yielding
141.75 meg per tape which is reasonable on a QIC-150 which can hold a maximum
of 150 meg.
 
michael pearlman
canuck@rice.edu (Mike Pearlman)
 
=============================================
 
I would like to thank all the people who have cared to answer my query.
I have tested the suggestions and I had no more problems.
 
Rubens Queiroz
Systems Support Analyst
State University of Campinas
email: queiroz@ccsun.unicamp.ansp.br
 
 
 



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:38 CDT