silver@s.cs.uiuc.edu (11/03/89)
Re: 6157-002 tape drives being slow Yes, I agree they aren't speed-demons, but I have obtained a tremendous speedup with buffering. The daily backups we make (5-8Meg) take only 4-5 minutes. The fulldumps (200+Meg) take only 75-80 minutes. That's not including the time to switch between the 4 tapes it requires. All we changed in the cpio (daily) or backup (fulldump) calls was to include -C2000 for a block/buffer size of 1 Meg. We have 12 Meg. of real memory, so page-faulting is not much of a problem. As far as reading -001 tapes on a -002, I'd like to answer to that if anyone ever discovers one. All of this is from a RT/125 with 3x310 Meg disks (mostly empty). Edward A. Silverstein silver@director.beckman.uiuc.edu
luner@werewolf.CS.WISC.EDU (David L. Luner) (11/04/89)
In article <213600004@s.cs.uiuc.edu> silver@s.cs.uiuc.edu writes: > >Re: 6157-002 tape drives being slow > >As far as reading -001 tapes on a -002, I'd like to answer to that >if anyone ever discovers one. If I recall correctly, and I'll check tonight, a -002 *can* read tapes written by a -001, but not the other way around. (And there is no way to make a -002 write tapes in such a way that a -001 can read them). -- David
ehrlich@cs.psu.edu (Daniel Ehrlich) (11/10/89)
In article <213600004@s.cs.uiuc.edu> silver@s.cs.uiuc.edu writes:
silver> Nf-ID: #R:nmtsun.nmt.edu:3425:s.cs.uiuc.edu:213600004:000:731
silver> Nf-From: s.cs.uiuc.edu!silver Nov 2 17:32:00 1989
silver>
silver> Re: 6157-002 tape drives being slow
silver>
silver> Yes, I agree they aren't speed-demons, but I have obtained a tremendous
silver> speedup with buffering. The daily backups we make (5-8Meg) take
silver> only 4-5 minutes. The fulldumps (200+Meg) take only 75-80 minutes.
silver> That's not including the time to switch between the 4 tapes it requires.
silver>
silver> All we changed in the cpio (daily) or backup (fulldump) calls was to
silver> include -C2000 for a block/buffer size of 1 Meg. We have 12 Meg. of
silver> real memory, so page-faulting is not much of a problem.
silver>
silver> As far as reading -001 tapes on a -002, I'd like to answer to that
silver> if anyone ever discovers one.
One could always read -001 tapes on a -002 drive. I think the original
question was about `writing' -001 tapes on the -002. Unfortunately the
answer would appear to be no. This is on AOS where I have looked at the
sources for the device driver. There is not even a hint ob being able to do
this.
silver> All of this is from a RT/125 with 3x310 Meg disks (mostly empty).
silver>
silver> Edward A. Silverstein
silver> silver@director.beckman.uiuc.edu
--
Dan Ehrlich <ehrlich@cs.psu.edu>
"A message is not a message until the rules for interpreting it
are in the hands of the receiver." --Apollo Belvedere Smithdarren@genieho.UUCP (Darren Holloway) (11/10/89)
In article <213600004@s.cs.uiuc.edu>, silver@s.cs.uiuc.edu writes: > > Re: 6157-002 tape drives being slow > > Yes, I agree they aren't speed-demons, but I have obtained a tremendous > speedup with buffering. The daily backups we make (5-8Meg) take > only 4-5 minutes. The fulldumps (200+Meg) take only 75-80 minutes. > That's not including the time to switch between the 4 tapes it requires. You can get about 10-20% more speed than -C2000 by using "dd", as such: To do a backup: li -Ra /u | backup -ivr -f - -C30 | dd of=/dev/rmt0 bs=30b To do a restore: dd if=/dev/rmt0 bs=30b | restore -xv -f - -C30 However, I have heard that it is possible to lose your end-of-tape markers while doing this...You will have to try it yourself and decide. -- ----------------------------------------------------------------------------- Darren G. Holloway Systems Administrator GENIE Computer Systems UUCP: calgary!xenlink!blender!genieho!{darren|root} REALTIME: +1 403 295 0418 "You'll get there Someday" - Slogan in Hyundai TV Commercial