don@seila.UUCP (Don Kossman) (03/29/89)
we have an Exabyte 8mm cartridge drive, with the Q-bus controller distributed by Contemporary Cybernetics, installed in a MicroVAX-II running Ultrix 3.0. We just had the drive upgraded to the latest ROM level. we use it for dumps, which work just fine, and have used it with tar as well, but only with the "c" option (ie create a new archive). when i use the tar "r" option (ie append to the end of the archive), however, i get an "error 0, writing to archive 1". the exact same operation works just fine with my TK50 and TU81-plus... i spoke to the vendor, and the tech support person thought he'd seen a similar problem in the past with tar and said that he'd check, but sort of thought that "tar r just doesn't work". meanwhile, i thought i'd ask the net.... has anybody seen this or something similar? IS THERE A WORKAROUND? (We don't have Ultrix source) TIA for any pointers or info. -- Don Kossman, SEI Information Technology, Los Angeles ...sun!tsunami!seila!don, -OR- mahendo.jpl.nasa.gov!jplgodo!seila!don
brian@ronin.cc.umich.edu (Brian Wolfe) (03/29/89)
In article <416@seila.UUCP> don@seila.UUCP (Don Kossman) writes: >we have an Exabyte 8mm cartridge drive, with the Q-bus >controller distributed by Contemporary Cybernetics, installed >in a MicroVAX-II running Ultrix 3.0. We just had the >drive upgraded to the latest ROM level. > >we use it for dumps, which work just fine, and have used >it with tar as well, but only with the "c" option (ie create >a new archive). when i use the tar "r" option (ie append to >the end of the archive), however, i get an "error 0, >writing to archive 1". the exact same operation works just >fine with my TK50 and TU81-plus... > >has anybody seen this or something similar? IS THERE A WORKAROUND? >(We don't have Ultrix source) Well I've experienced a few bits of weirdness with a TTi CTS-8 (I think the drive mechanism is actually made by Exabyte too) and tar. You might try using mt to position to the end of tape and then try a tar c (I haven't actually tried this) using /dev/nrmt# to keep appending archives. Its nasty, but it may be worth trying. I frequently get archive blocksize=0 errors on my 8mm when I try to read from it, the mt fsf sometimes does the trick. ------------------------------------------------------------------------- Brian Wolfe Internet: BAW@um.cc.umich.edu Systems Analyst UUCP: {ucbvax,uunet}!umix!hfhrv!brian Henry Ford Hospital Detroit MI 48202 (313)-876-7461 -------------------------------------------------------------------------
grr@cbmvax.UUCP (George Robbins) (03/29/89)
In article <416@seila.UUCP> don@seila.UUCP (Don Kossman) writes: > we have an Exabyte 8mm cartridge drive, with the Q-bus... > > we use it for dumps, which work just fine, and have used > it with tar as well, but only with the "c" option (ie create > a new archive). when i use the tar "r" option (ie append to > the end of the archive), however, i get an "error 0, > writing to archive 1". the exact same operation works just > fine with my TK50 and TU81-plus... The tar -r option probably requires that the device implement some approximation of the traditional magnetic tape "skip" functions. It is quite possible that this device can't emulate these function. If you are trying to use a single tape over and over while appending new data to the end of the prior archive, you may be SOL. You might be able to "read" to the end of the tar "file" and then write a complete new archive, but operationally you'ed probably end up writing over the wrong file a bit too often. Hopefully, you write multiple tar archives to a single tape, simply by using a "no-rewind" device name, as long as you don't have to remove the tape from the unit. How does it work otherwise? I'm thinking about getting one of these toys - multi-reel backups @6250 BPI are getting to be a pain. -- George Robbins - now working for, uucp: {uunet|pyramid|rutgers}!cbmvax!grr but no way officially representing arpa: cbmvax!grr@uunet.uu.net Commodore, Engineering Department fone: 215-431-9255 (only by moonlite)
maj@cl.cam.ac.uk (Martyn Johnson) (03/30/89)
> i spoke to the vendor, and the tech support person thought he'd > seen a similar problem in the past with tar and said that he'd check, > but sort of thought that "tar r just doesn't work". "tar r" CAN'T work on an Exabyte. It involves reading the last block of the archive and rewriting it. Exabytes can't do that, because the data blocks written in a diagonal stripe along quite a long length of tape, whereas the erase head is vertical and erases the full width to tape. It is therefore impossible to erase isolated data blocks. Incidentally, this also explains why tapemarks eat up so much of the tape capacity. Martyn Johnson maj%uk.ac.cam.cl@nss.cs.ucl.ac.uk University of Cambridge Computer Laboratory