rogerj@batcomputer.tn.cornell.edu (Roger Jagoda) (07/16/90)
Folks, I have inherited a UNIX system (VAX/ULTRIX) with a SCSI Exabyte (OEM'd 3rd Party as usual) Tape unit. My question regards the parameters for the best "dump" command use. For example, we're using now: dump <level>usdf 4800 6666 /dev/nrmt1h <f. system> This assumes a density or 4800 and a tape length of 6666 feet which CAN't be correct. Something is wrong as we can't restore files either..:-( The tape unit responds with "not a dump tape, I/O error" So, the tapes are SONY 120Ps and the ExaByte manual says the drive is capable of writing "linear recording density = 43200 bits/inch" Now, isn't this the number we should use instead of 4800 (which sounds more like a guess from the old 1600 bpi tape reel days..)? As for feet of tape length, we have asked everyone including the mail order house where the tapes were bought, NO ONE seesm to know how many feet of tape come on a 120 tape! The actual model number of the tape is "P6-120MP" for "Metal Particle". If anyone knows how long these tapes are can they post an answer? Thanks. Actually, couldn't we just put a very large number for the lenght (like 100000)? That number is just so the tape doesn't run off the reel during the dump, right? Well, our dumps don't take up 2.2G so there's probably no danger of that. Should our dump command be more like: dump <level>udsf 43200 100000 /dev/nrmt1h <f.system> ? Thanks in advance for any pointers or comments! --Roger Jagoda --Cornell University -- ------------------------------------------------------------------------- Roger Jagoda -- My employers don't even like paying Cornell University me, let alone accept responsibility fqoj@cornella.cit.cornell.edu for anything I say or do! --
merlyn@iwarp.intel.com (Randal Schwartz) (07/16/90)
In article <10542@batcomputer.tn.cornell.edu>, rogerj@batcomputer (Roger Jagoda) writes: | Should our dump command be more like: | | dump <level>udsf 43200 100000 /dev/nrmt1h <f.system> ? Here's what my vendor told me to use: dump <level>dsbf 43000 12000 64 /dev/nrst1 <filesystem> The "64" can be any multiple of 8 from 8 to 120, but 64 is easy to remember and document. This commandline gives sensible figures for portions of a tape used (204.64MB on 0.09 tapes, for example, said a recent dump). Yes, you could use a huge figure for length as well, unless you plan on having a 2.2GB filesystem. :-) I've been running with a commandline like this for about eight months from 13 microvaxen and 40 sun3, sun4, and sun386 machines. No problem, as Alf would say. Just another exabyte hacker, -- /=Randal L. Schwartz, Stonehenge Consulting Services (503)777-0095 ==========\ | on contract to Intel's iWarp project, Beaverton, Oregon, USA, Sol III | | merlyn@iwarp.intel.com ...!any-MX-mailer-like-uunet!iwarp.intel.com!merlyn | \=Cute Quote: "Welcome to Portland, Oregon, home of the California Raisins!"=/
samlb@pioneer.arc.nasa.gov (Sam Bassett RCS) (07/17/90)
Pioneer is a VAX 11/785 running Ultrix 3.0, and we're having quite good luck with specifying a density of 1600 and a length of 121845 for our backups to a Micro Technology Inc Exabyte drive -- no problems writing or reading. Sam'l Bassett, Sterling Software @ NASA Ames Research Center, Moffett Field CA 94035 Work: (415) 604-4792; Home: (415) 969-2644 samlb@well.sf.ca.us samlb@ames.arc.nasa.gov <Disclaimer> := 'Sterling doesn't _have_ opinions -- much less NASA!'
davec@pyra.co.uk (David Collins) (07/17/90)
In article <10542@batcomputer.tn.cornell.edu> rogerj@tcgould.tn.cornell.edu (Roger Jagoda) writes: >Folks, > >I have inherited a UNIX system (VAX/ULTRIX) with a SCSI Exabyte (OEM'd 3rd Party as usual) Tape unit. My question regards the parameters for the best "dump" >command use. For example, we're using now: > >dump <level>usdf 4800 6666 /dev/nrmt1h <f. system> > >This assumes a density or 4800 and a tape length of 6666 feet which CAN't >be correct. Something is wrong as we can't restore files either..:-( > >The tape unit responds with "not a dump tape, I/O error" > >So, the tapes are SONY 120Ps and the ExaByte manual says the drive is capable >of writing "linear recording density = 43200 bits/inch" > >Now, isn't this the number we should use instead of 4800 (which sounds more >like a guess from the old 1600 bpi tape reel days..)? > >As for feet of tape length, we have asked everyone including the mail >order house where the tapes were bought, NO ONE seesm to know how many >feet of tape come on a 120 tape! The actual model number of the tape is >"P6-120MP" for "Metal Particle". If anyone knows how long these tapes >are can they post an answer? Thanks. Actually, couldn't we just put a >very large number for the lenght (like 100000)? That number is just >so the tape doesn't run off the reel during the dump, right? Well, our >dumps don't take up 2.2G so there's probably no danger of that. > > >Should our dump command be more like: > >dump <level>udsf 43200 100000 /dev/nrmt1h <f.system> ? > >Thanks in advance for any pointers or comments! > >--Roger Jagoda >--Cornell University > >-- >------------------------------------------------------------------------- >Roger Jagoda -- My employers don't even like paying >Cornell University me, let alone accept responsibility >fqoj@cornella.cit.cornell.edu for anything I say or do! -- I think that you will find that the density that the exabyte simulates is 1600 and that if you are using the "90" minute tapes the length of the tape is 116000 feet. Those are the parameters (and tapes) that I use on our inhouse systems and I havent had a problem when needing to restore files "that disappeared". If you want the equation to work out the length: The 90 minutes (standard play) tape can hold over 2.0 gigabytes. (Block size) * (Tape Length) 0.95 * 2143303680 = ---------------------------- = Tape Capacity = 2036138496 (Block size) ------------ + 0.6 1600 Where : Block size = 2048 * 5 = 10k (default "dump" parameters) Tape length = the length of the tape in inches ("dump" requires feet). 1600 = B.P.I. density emulated by the drive. 0.6 = Inter record gap (inches). The tape length calculated from this formula is 116000 ft. The formula above originated from Mike Shrimpton at Interface Distributors Ltd and as I stated above the results have served me alright. BTW If you are backing up across an ethernet a block size of 32 appears to speed up the dumps on our inhouse systems. Good Luck -m------- Dave Collins email : davec@uk.co.pyra ---mmm----- Pyramid Technology Ltd fax : (+44) 252 373135 -----mmmmm--- Pyramid House, Solartron Road voice :work (+44) 252 373035 -------mmmmmmm- Farnborough, Hants, GU14 7PL UK :home (+44) 635 72044 "I turn to my computer as my friend" Kate Bush - from Deeper Understanding
rick@oliveb.OLIVETTI.COM (Rick Meneely) (07/18/90)
Roger Jagoda: > I have inherited a UNIX system (VAX/ULTRIX) with a SCSI Exabyte (OEM'd 3rd Party as usual) Tape unit. My question regards the parameters for the best "dump" > command use. For example, we're using now: > > dump <level>usdf 4800 6666 /dev/nrmt1h <f. system> My Exebyte manual for an EXB-8200 from Perfect Byte, Inc. suggests using 54000 bpi and 6000 ft for a 120 minute cartridge tape - for 'dump'. It notes that these are not "real" values but are "faked" in order to get 'dump' to work properly. In essence 'dump' outsmarts itself. It also notes that if you are writing multiple dumps to a single tape you should decrease the size (length) parameter proportionally for sucessive dumps. > As for feet of tape length, we have asked everyone including the mail > order house where the tapes were bought, NO ONE seesm to know how many > feet of tape come on a 120 tape! The actual model number of the tape is > "P6-120MP" for "Metal Particle". If anyone knows how long these tapes > are can they post an answer? 106 Meters =~ 344 ft. (seems small doesn't it!) - Rick Meneely ------------------------------------------------------------------------- Disclaimer: * .* Internet: rick@olivee.ATC.Olivetti.Com Olivetti Advanced Technology Center
hans@smab.se (Hans C Larsson) (07/21/90)
I think tape speed is something like 2 cm/sec with a *PAL* Video8. Beware that: NTSC PAL ============== P6-120 = P5-90 (this is why there is no P5-120...) P6-90 = P5-60 etc. since NTSC-VCR's runs slower. -- ------------------------------------------------------------ Hans C Larsson Email: hans@smab.se Saab Missiles, Sweden