scott@rdahp.UUCP (Scott Hammond) (06/13/89)
System: HP9000/300 running HPUX 6.2, with HP9144 Cartridge tape drive I noticed something unusual last week. I was trying the -S 8 option to tcio which specifies a buffer size for talking to the tape drive which lets the computer keep up with the tape drive so it can stream (crummy explanation but you get the idea, note this only works for certain tape drives). Upon trying it, I discovered it only worked for 3M HP-compatible tapes (DC 600HC), and did not work with HP tapes. This also shows up with mediainit. I tried forcing a mediainit on a 600' HP tape, and it took well over 3 hours. The same on a 3M 600' tape required about an hour. The mediainit on the HP tape did not help matters when I retried writing to the tape. So, what is happening? Is the low-level format that different? I don't know much about tape formatting, but I get the impression that each leg of the serpentine track reaches from one end of the tape to the other on the 3M tape, but is broken up into much smaller chunks on the HP tape. Does this sound reasonable? Am I doing it wrong? Is this machine capable of reformatting cartridge tapes? (and of course, if so, how?). Have I gotten hold of some old HP tapes? (I didn't buy them, they were received from another site). Thanks for any and all information. -- Scott Hammond, R & D Associates, Marina del Rey, CA (213) 822-1715 : {tah386,randvax,ladcgw,<site-with-router>}!rdahp!scott : scott@harris.cis.ksu.edu
maddog@cbnews.ATT.COM (john.j.tupper) (06/13/89)
In article <234@rdahp.UUCP> scott@rdahp.UUCP (Scott Hammond) writes: >System: HP9000/300 running HPUX 6.2, with HP9144 Cartridge tape drive > >I noticed something unusual last week. I was trying the -S 8 option to >tcio which specifies a buffer size for talking to the tape drive which >lets the computer keep up with the tape drive so it can stream (crummy >explanation but you get the idea, note this only works for certain tape >drives). Upon trying it, I discovered it only worked for 3M >HP-compatible tapes (DC 600HC), and did not work with HP tapes. This >also shows up with mediainit. I tried forcing a mediainit on a 600' HP >tape, and it took well over 3 hours. The same on a 3M 600' tape >required about an hour. Each tape is formatted in blocks. The mediainit program goes out checking for bad blocks and will use spare blocks at the end of the tape as spares (there's a pool of spare blocks saved for just this purpose). My guess is that the hp tape you have has some of bad blocks. [A technicality: I'm not sure mediainit does the sparing of bad blocks, but something sure does.] We also have both hp and 3M tapes and have not noticed any difference between the two (except for the price :-). maddog ------------------------------------------------------------------------- sdlkf oiwoi sdf My real signature is illegible too
rjn@hpfcdc.HP.COM (Bob Niland) (06/13/89)
re: "I was trying the -S 8 option .... so it can stream..." > Upon trying it, I discovered it only worked for 3M > HP-compatible tapes (DC 600HC), and did not work with HP tapes. > This also shows up with mediainit. I tried forcing a mediainit on a 600' HP > tape, and it took well over 3 hours. The same on a 3M 600' tape required > about an hour. > The mediainit on the HP tape did not help matters when I retried writing > to the tape. I have never seen a re-certify (mediainit -r) take over an hour on an HP- supplied tape. The behaviour you describe suggests a defective tape. The extra two hours is due to excessive sparing of bad blocks. Please contact your cartridge tape source (dealer or HP sales office) and ask them how to proceed. They will probably want whatever historical information you have on those tapes. For example, has the tape ever been exposed to a high magnetic field. Incidentally, HP buys most (if not all) of its DC600HC, DC615HC, DC600XTD and DC615XTD tapes from 3M. Although many vendors make DC600-style cartridges, only 3M or a licensee is equipped to write the full-track "keys" that define the blocks on an HCD-format tape. 3M only recently started licensing, and I'm not sure if HP has purchased any tapes from a second source. The HP 88140LC, 88140SC media are pre-certified to HP specs. Generic DC6xxHC and DC6xxXTD tapes are not. 'mediainit -r' should actually take less time on an "HP" tape than a 3M tape. Regards, Hewlett-Packard Bob Niland rjn%hpfcrjn@hplabs.HP.COM 3404 East Harmony Road [hplabs|hpu...!hpfcse]!rjn Ft Collins CO 80525-9599
jewett@hpl-opus.HP.COM (Bob Jewett) (06/15/89)
> I have never seen a re-certify (mediainit -r) take over an hour on an HP- > supplied tape. The behaviour you describe suggests a defective tape. The > extra two hours is due to excessive sparing of bad blocks. We have also seen several tapes that were very slow. The problem seems to be confined to certain date codes. On the back of the tape (usually in dark blue ink) is the date code, which starts something like: 5A832 ^^^ YWW (year and week) The bad range is week 32 through week 43 of year 8 (1988). The symptom is that data takes MUCH longer to record. I check our tapes in that date range by timing the write of a one meg file. Bob Jewett
cjames@hpldsla.HP.COM (Craig James) (06/15/89)
Scott Hammond writes: > System: HP9000/300 running HPUX 6.2, with HP9144 Cartridge tape drive > I noticed something unusual last week. I was trying the -S 8 option to > tcio which ... lets the computer keep up with the tape drive so it can > stream ... Upon trying it, I discovered it only worked for 3M > HP-compatible tapes (DC 600HC), and did not work with HP tapes. There is a small chance that your tape is from a bad batch. A year or two ago HP shipped a few that were formatted wrong, screwing up the adjacency of various sectors in a way I don't pretend to understand. These tapes take many, many times longer to do things than good ones. Perhaps someone else at HP recalls the date codes of these tapes? Craig James, HP Labs