david@pyr.gatech.EDU (David Brown) (06/02/89)
Hello, all. I hate to take everyone away from the GNU kernal discussion, but... A short while ago, someone put some very important drawings onto a 1/4" tape using tar. Subsequently, and end-of-tape marker was written at the beginning of the tape. I think that trashed the first drawing. but the others should still be around, just not readily accessable (correct me if I'm wrong). Is there a PD or standard UN*X utility that will read the info that's past the end-of-tape marker? I'm running SunOS4.0 on a 3/60. The tape was written under SunOS3.5. Please respond via e-mail. Thanks for any help. david ---------------------------------------------------------------------------- David Brown Armstrong State College, Savannah, Georgia uucp: ...!gatech!gitpyr!david ARPA: david@pyr.gatech.edu
samperi@marob.masa.com (Dominick Samperi) (06/03/89)
In article <8413@pyr.gatech.EDU> david@pyr.gatech.edu (David Brown) writes: >A short while ago, someone put some very important drawings >onto a 1/4" tape using tar. Subsequently, and end-of-tape >marker was written at the beginning of the tape. I think >that trashed the first drawing. but the others should still >be around, just not readily accessable (correct me if I'm >wrong). Is there a PD or standard UN*X utility that will >read the info that's past the end-of-tape marker? No EOT marker is placed after the last file on streaming 1/4" mag tape. Each file (or segment) is terminated with a file-mark, and there is an end-of-media hole at the physical end of tape. The end of readable data is located by reading file-marks until a "No Data Detected" error is returned from the controller. Data can be appended after this state is reached, but I do not think these drives/controllers can resync and start reading blocks after the "No Data Detected" point. One reason for this is the fact that these "intelligent" drives do error detection by keeping track of the block sequence number, and they refuse to pass data to the host if blocks are not read in the correct order. For this reason, all software that is written for these drives MUST detect when the user is attempting to overwrite existing data, and issue appropriate warnings before continuing. readable data on the tape by reading file-marks until a "No Data Detected" -- Dominick Samperi -- ESCC samperi@marob.masa.com uunet!hombre!samperi
buck@siswat.UUCP (A. Lester Buck) (06/04/89)
In article <8413@pyr.gatech.EDU> david@pyr.gatech.edu (David Brown) writes: >A short while ago, someone put some very important drawings >onto a 1/4" tape using tar. Subsequently, and end-of-tape >marker was written at the beginning of the tape. I think >that trashed the first drawing. but the others should still >be around, just not readily accessable (correct me if I'm >wrong). Is there a PD or standard UN*X utility that will >read the info that's past the end-of-tape marker? [I guess this was posted separately to comp.periphs, where I posted a longer reply.] The QIC-24 standard requires 45 inches of blank (erased) tape to follow the last data recorded, and when writing track 0, the controller erases all 9 tracks at once. It would take a very specialized driver violating the standard to read such a tape. -- A. Lester Buck ...!texbell!moray!siswat!buck