baldwin@usceast.UUCP (Chuck Baldwin) (06/14/88)
Hi All: Several weeks ago, I posted a request to the net for information concerning Helical-Scan backup devices. Of the five, count'em 5, responses I received three were requests for a summary of the information I obtained. Well, as you can guess, there aren't many in use. The only reply from someone with a Helical-Scan backup device in use follows: > We got our "Exabyte 8mm Tape Drive Subsystem" for our Sun from >Perfect Byte, Inc, 7121 Cass Street, Omaha, Nebraska 68132, 402- >554-1122. Keith Wagner is the president (ihnp4!pbi!keith). The >unit came with the Exabyte 8200 drive, a Ciprico 3500 controller, >cables, and software. It was easy to install and preforms the >standard Unix tape i/o just fine. 'Tar', 'dd', 'dump/restore' have >all been tested and found to work. I can 'dump' at about 10Mb per >minute to the drive. Should be able to 'dump' at least 4 Eagle XPs >to one tape. Tapes cost about $8 each. . . . > bull@noao.arizona.edu > Frank Bull > Cadet Commander, Solar Patrol (Sun Systems Administrator ;-) > Central Computer Services > National Optical Astronomy Observatories . . . The following reply, from David Lesher, has created a lot of rethinking about the entire problem: >I am very wary of such devices. Since there is NO way to >verify (read after write) on a vcr [besides stopping the tape, >going back to the beginning, reading and comparing, etc] >the ones I have seen record each file several times in hope >that one is perfect. >Do you want to risk YOUR data to such a bet? >On the other hand, the WORMS I have seen look nice. For those who don't know, WORMS, stands for Write Once Read Many. The most readly seen WORM is a laser disk, i.e. CD player. These storage devices store HUGE amounts of data, but only once. This may be useful for system backups, or possibly everyday backups depending on how much data is being stored. Also, I have been reading comp.sys.sun and the SunSpots literature about 8mm Helical-Scan devices on Suns; and there is some discussion of the various error recovery schemes in the devices. Anyway, we're examining the cost effectiveness and error recovery inherent in this solution. Considering the response we have received we're still not going to make a purchase soon. In the mean time we are going to keep on worrying with 2400' tape. Thanks again for all the help and if anyone gets one of these devices please drop me a note about your experiences. Chuck Baldwin -- =============================================================================== UUCP: ...seismo!ncr-sd!ncrcae!usceast!baldwin CSNET: baldwin@cs.scarolina.edu Chuck Baldwin, Dept of Math, University of South Carolina, Columbia, SC 29208 ===============================================================================
blarson@skat.usc.edu (Bob Larson) (06/14/88)
In article <2648@usceast.UUCP> baldwin@usceast.UUCP (Chuck Baldwin) writes: > The following reply, from David Lesher: >>there is NO way to >>verify (read after write) on a vcr [besides stopping the tape, >>going back to the beginning, reading and comparing, etc] While this is true of your ordinary VCR, I've heard it is not true of at least some computer tape drives that use VCR tapes. There is no techincal reason on a VHS style tape that prevents extra tape from beeing pulled out of the cartrage to run it over a second rotaing tape head. The 4 mm (DAT) cartrage seems based on the same type of technology. (I've never examined an 8 mm tape closely.) This obviously wouldn't be cheap, but with prices more than an order of magnatude higher than home VCRs, it seems beleivable. [I've been looking at three vendors with alternative tape solutions for Primes, based on VHS, 8mm, and 4mm(DAT). (Arn't standards lovely? Tapes should be easily available, but nobody can read tapes from the other brands.)] >>the ones I have seen record each file several times in hope >>that one is perfect. Early ones (and probably the current cheepos) did do this, but hopefully there has been some advance in the last 5 years. -- Bob Larson Arpa: Blarson@Ecla.Usc.Edu blarson@skat.usc.edu Uucp: {sdcrdcf,cit-vax}!oberon!skat!blarson Prime mailing list: info-prime-request%fns1@ecla.usc.edu oberon!fns1!info-prime-request
chris@mimsy.UUCP (Chris Torek) (06/14/88)
In article <2648@usceast.UUCP> baldwin@usceast.UUCP (Chuck Baldwin) writes: > Several weeks ago, I posted a request to the net for information >concerning Helical-Scan backup devices. ... > The following reply, from David Lesher, has created a lot of rethinking >about the entire problem: >>I am very wary of such devices. Since there is NO way to >>verify (read after write) on a vcr [besides stopping the tape, >>going back to the beginning, reading and comparing, etc] >>the ones I have seen record each file several times in hope >>that one is perfect. ... Note that the Exabyte 8mm videotape backup (and another similar 8mm system---I forget whose, but it has a Pertec interface instead of SCSI; it is the same drive with a different controller) are not VCR backup systems. Apparently most VCR systems do indeed have no motor control. The Exabyte does, hence it could do verification. (Whether it in fact does, I have no idea.) -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris
wyatt@cfa.harvard.EDU (Bill Wyatt) (06/14/88)
>> Several weeks ago, I posted a request to the net for information >>concerning Helical-Scan backup devices. ... > >>>I am very wary of such devices. Since there is NO way to >>>verify (read after write) on a vcr [besides stopping the tape, [...] > > Note that the Exabyte 8mm videotape backup (and another similar 8mm > system---I forget whose, but it has a Pertec interface instead of SCSI; > it is the same drive with a different controller) are not VCR backup > systems. Apparently most VCR systems do indeed have no motor control. > The Exabyte does, hence it could do verification. (Whether it in fact > does, I have no idea.) I missed the original request, or I would have supplied some info here. The Exabyte 8mm system uses a special head that DOES allow read-after-write. Thus, blocks in error are caught and rewritten right away, some number of physical blocks down the tape. Combined with about a 40% ECC overhead (i.e. about 400 extra bytes are written for each 1024 data bytes), the stated error rate of 1 in 10^13 is quite reasonable. If you want a few hundred word summary of our experiences with these (we've had a few for several months), e-mail me and I'll respond. -- Bill UUCP: {husc6,ihnp4,cmcl2,mit-eddie}!harvard!cfa!wyatt Wyatt ARPA: wyatt@cfa.harvard.edu (or) wyatt%cfa@harvard.harvard.edu BITNET: wyatt@cfa2 SPAN: cfairt::wyatt
terryk@encore.UUCP (Terence Kelleher) (06/14/88)
The Exabyte drive has both read and write heads on the same drum and does do a read after write. I would assume that the 4mm DAT tape drives will also read after write. Terry Kelleher
bull@noao.arizona.edu (Frank Bull CCS) (06/14/88)
In article <2648@usceast.UUCP> baldwin@usceast.UUCP (Chuck Baldwin) writes: > The following reply, from David Lesher: >>there is NO way to >>verify (read after write) on a vcr [besides stopping the tape, >>going back to the beginning, reading and comparing, etc] From the Product Specification manual, EXB-8200 8MM Cartridge Tape Subsystem, Exabyte Corp. "During WRITE operations, the EXB-8200 CTS records servo information and formatted user data blocks on the same track and performs a read-after- write check of the recorded user data. In the event of an error during the read-back check for a WRITE error condition, error recovery procedures are performed without host intervention, and without the need to reposition the tape." The "Read" head is positioned 180 degrees from the "Write" head and does a read-back-check during the 2nd 180 degrees of head rotation.
johnm@voltron.SGI.COM (Foosball Addict) (06/22/88)
In article <1044@cfa.cfa.harvard.EDU>, wyatt@cfa.harvard.EDU (Bill Wyatt) writes: > The Exabyte 8mm system uses a special head that DOES allow > read-after-write. Thus, blocks in error are caught and rewritten > right away, some number of physical blocks down the tape. ^^^^^^^^^^ The Exabyte drive does a read after write to compare and if there are any errors, it re-writes the date 11 blocks away. The reasoning is that typical media flaws are horizontal in nature so the block the data is re-written to is skewed off the line the error was detected. johnm