randyh@ozvax.GWD.TEK.COM (Randy Hendry) (08/10/88)
I'm having problems getting Amiga partitions recognized and formatted on
my bridgeboard harddisks. Hopefully someone will have some insight and
solutions to all the errors I have encountered. My configuration is
Computer:
Amiga 2000 Rev 4.2, small function keys on keyboard
Bridgeboard (ibm xt) Assy No. 380755 Rev. 3 in right (from front) slot
380788-04 EMu-BIOS 3.4 (Phoenix), 38078[345]-01 support chips
Harddisk:
bridge: Western Digital controller WD1002A-WX1
ROM = 62-000094-002 (latest SuperBIOS), in middle XT slot
Drive 1: Maxtor XT-1105 (918 cyl, 11 heads, ~80MB)
Drive 2: Micropolis 1304 (830 cyl, 6 heads, ~40MB)
Software:
Amiga: WB 1.2 (33.59) w/ bridgeboard support (version ??? (~Sept 87))
bridge: MS-DOS 3.3 from MS + adisk, etc from C='s MS-DOS 3.2
The partitions I'm trying to set up are as follows:
Drive 1: cyl's 0 - 56, MSDOS3.3 C: ~5MB (system disk. works fine)
cyl's 7 - 113, MSDOS D: ~5MB (formatted fine (empty))
cyl's 114 - 343, Amiga jh0: ~20MB (formatted fine (empty))
cyl's 344 - 917, currently unused, planned mostly for Amiga
Drive 2: cyl's 0 - 310, Minix ~15MB (current not partitioned)
cyl's 311 - 829, Amiga jh1: ~25MB (THIS IS THE ONE THAT FAILED)
Sorry about all the detail, but better too much than too little. Anyway, my
scenario is as follows: Low level format both drives using the HD controller's
SuperBIOS. Only enter the worst of the bad spots into the bad track table
(btw, I have tried it w/ all the original known bad areas). Fdisk and high
level format the MS-DOS partitions (always rebooting as necessary). Adisk
both amiga partitions. Djmount and cancel requestors (two of them both
saying something like "non-DOS" disk (as expected)).
"dpformat drive jh0: name jh0"; works fine.
"dpformat drive jh1: name jh1"; When I return shows this on the screen:
Format cyl 516. Retry required cylinder 1. Hard error cylinder 1.
Error: Format abandoned. A requester also appeared (I'm guessing an
after product of the unsuccessful format): Error validating disk. Key
48340 checksum error.
This is the first problem. Questions:
1. Did the format find a bad area on the disk?
2. If so, how can I tell where it was so I can lock it out with the low-level
formatting. (Cyl 1 isn't very helpful since I don't want to waste that
much space).
3. Will it do any good to lock out such bad spots i.e. will the driver
handle them. My guess is yes, since I (think) I tried this before.
4. Is there some problem w/ the size (~25MB) or the position of the partition?
Now in trying to get all this to work, I decide to use the same partition (at
least the first 5 cylinders where the problem was indicated) as an MS-DOS
partition to see if it will format. After making all the changes using
adisk and fdisk, I was able to format the area successfully. Then, just
for kicks :-( I decide to try and format it under the Amiga side again.
Again I make judicious use of fdisk and adisk, but after rebooting, when
I do djmount it doesn't recognize jh1: as a valid hard disk partition (my
guess). Jh0: still comes up fine, but instead of getting a requestor
saying something like "non-dos disk", I get one saying "Please insert
volume JH1 in any drive"! It's like parts of the OS now think jh1: is a floppy.
Well, getting desperate, I shut down drive 1 and reboot. This time djmount
just hangs with an interesting "Divide by zero" error that shows up on, of
all places, the MS-DOS screen.
Thus the second problem. Questions:
1. Why does djmount no longer recognize my second hard disk partition?
(Is this what C= means when, in the pamplet that comes with the
bridgeboard, it is recommended that all partitions be formatted
at once?)
2. Will I not be able to format additional partitions later on
when I decide to use the unused space on my 80MB drive?
3. Is any of this detailed in the Tech. Ref. which I have on order?
My local dealer isn't much help, *he* thinks there's an RLL
harddisk controller on the bridgecard...
4. The big question: what do I need to do to get all this to work?
I'm at witt's end (and don't even have a magazine :-) I sure could use
some help from someone who has successfully made it through this
procedure.
Thanks,
Randy
--
Randy Hendry randyh@ozvax.gwd.tek.com ...!tektronix!ozvax!randyh
Tektronix, Inc. Graphic Terminals Division (503)685-3063w
P.O. Box 1000, M.S. 63-523, Wilsonville, OR 97070 (503)639-7697h
"In Oregon, the sun shines every day; in the summer, you can even see it."brianr@tekig4.TEK.COM (Brian Rhodefer) (08/10/88)
Last week, I had almost exactly the same experience that Randy Hendry has so calmly described: inability to install and format an 80Mbyte harddisk as JH0: on my Bridgeboard. Same model drive, same model HD controller. I didn't post to the net at the time, partly because I haven't e-mailed my thanks to the folks who replied to my LAST two questions, partly because it took me TWO DAYS to get my amiga working again, but mostly because I was afraid of what I'd say in my enraged state. News that Janus has claimed another victim, however, cannot be borne in silence. My interpretation of the machine's behavior (and please, oh please God let me be wrong) is that the `dpformat' software (which, I want to point out, is intended SPECIFICALLY to format a hard disk), (*>hysterical, outraged, purple-faced-shrieking mode ON <*) CAN'T TOLERATE **ANY** FLAWS ON THE DRIVE! IT DOESN'T MAP-OUT FLAWED TRACKS, it MAPS-OUT FLAWED DRIVES!!!! Is this some sort of cruel bait-and-switch tactic? Am I just learning my lesson for not shelling out another grand for an Amiga-side hard disk? Even IBM & Microsoft, who still brazenly ship that abominable `Edlin' editor with their MS-DOS, know how to supply a functional disk-formatter. [attempt to return to civility:] The requestor that popped up 40 minutes into the `dpformat' suggested that I use `DISKDOCTOR' to un-corrupt the disk. I did. Diskdoctor printed out a list of disk hard errors, and then advised that the disk be re-formatted after saving any of its files. Since JH0: was blank, I proceeded to re-format, again with `dpformat'. Forty (though it would've taken the Bridgeboard's 8088 only eight) minutes later, the exact same requestor, and the exact same nonusable JH0:. I don't know whether I've guessed at the correct solution, but what I've done is to re-run `DISKDOCTOR', and then used the JH0: drive without subsequently re-formatting it. I'm hoping that DISKDOCTOR did the equivalent of flaw-mapping. Side note: Gee, it would've been nice if C-A had provided documentation to explain HOW to use `diskdoctor', particularly when the built-in system requestors instruct people to use it. Everyone, I suppose, is expected to become Registered Developers? Why not include it in the system purchase price? Still waving the Amiga flag, but not as enthusiastically, Brian Rhodefer ...!tektronix!tekig4!brianr
andy@cbmvax.UUCP (Andy Finkel) (08/12/88)
In article <3090@tekig4.TEK.COM> brianr@tekig4.TEK.COM (Brian Rhodefer) writes: >Last week, I had almost exactly the same experience that Randy Hendry >has so calmly described: inability to install and format an 80Mbyte >harddisk as JH0: on my Bridgeboard. Same model drive, same model HD >controller. > AmigaDOS V1.2 only supports about 54meg per partition in V1.2. That's what your problem is. Make your partition smaller, and things will work better. This limit is removed under 1.3, using the FastFileSystem. andy > CAN'T TOLERATE **ANY** FLAWS ON THE DRIVE! IT DOESN'T > MAP-OUT FLAWED TRACKS, it MAPS-OUT FLAWED DRIVES!!!! > Actually, it uses the exact same bad track method that is used on the PC side for all of its XT hard disks. Since this wasn't your problem, it didn't kick in. >The requestor that popped up 40 minutes into the `dpformat' suggested >that I use `DISKDOCTOR' to un-corrupt the disk. I did. Diskdoctor AmigaDOS did some sanity checks, didn't like what it saw, and advised you to recover your files using its dos level recovery program. (its not a bad block mapper) >printed out a list of disk hard errors, and then advised that the >disk be re-formatted after saving any of its files. Since JH0: was >blank, I proceeded to re-format, again with `dpformat'. >Forty (though it would've taken the Bridgeboard's 8088 only eight) >minutes later, the exact same requestor, and the exact same nonusable JH0:. For an 80 Meg drive ? Hmmm, I thought PC's could only go up to 32 :-) DPFormat was writing data and reading it back, to each track; that does take awhile. Especially since its sending it through the Bridge link, rather than calling whatever bios rom routine the PC usually calls. >Everyone, I suppose, is expected to become Registered Developers? Nope; not everyone qualifies. Some docs are in the Bantam AmigaDOS manual; we are including a bit more in our 1.3 manual. Sorry you had problems; also, sorry you spent 2 days steaming; however, if your still hot now, I guess it was the right idea :-) andy -- andy finkel {uunet|rutgers|amiga}!cbmvax!andy Commodore-Amiga, Inc. "If we can't fix it, it ain't broke." Any expressed opinions are mine; but feel free to share. I disclaim all responsibilities, all shapes, all sizes, all colors.
randyh@ozvax.GWD.TEK.COM (Randy Hendry) (08/16/88)
The saga continues... I wrote: >Why does djmount no longer recognize my second hard disk partition? I seemed to have answered this one myself (at least partially). (If you remember, some part of the OS was refusing to recognize jh1: as a hard disk partition.) The problem seems to be related to me making the partition *too small*. I had made it only 5 cylinders (~0.25MB) for testing when I got the strange failure. When I increased it up to 2MB (~40 cylinders) djmount, et al. worked fine again. My conclusion is that there is some lower limit for a disk partition size that I exceeded. Anyway, onward... (i.e., now I'm back to the original problem:) >"dpformat drive jh1: name jh1"; >Format cyl 516. Retry required cylinder 1. Hard error cylinder 1. Brian Rhodefer (brianr@tekig4.TEK.COM) suggested that I try DiskDoctor on the sick partition. Good idea; I did. When I ran DiskDoctor it, started reporting: (paraphrased) Hard Error Track 1 Surface 6. ... Hard Error Track 1 Surface 10. Hard Error Track 2 Surface 6. ... And did so for every track (cylinder) of the disk (until I stopped it that is). There is a good reason for this: There are no heads (surfaces) 6-10 on this disk! The disk has 6 heads (0-5). Now non-coincidentally :-( the disk on which partition jh0: resides *does* have 11 heads (0-10). It appears that djmount (or dpformat) is using the head information for the first drive for all drives. BTW, I thought I might have things hooked up wrong, but MS-DOS doesn't have a problem with the drives and also David Fisk <dfisk@hawk.ulowell.edu> wrote to me saying that he was having the exact same problem and even came to same conclusion about djmount(?) assuming the wrong number of heads. Seeing this error made me think that I might be able to tell djmount manually about the number of heads by using MountList. I tried a couple of combinations, but had no luck. Djmount doesn't seem to use MountList -- is this true? So now, assuming my problem is truly with the number of heads, 1. What do I do to get Janus to know about the difference in the drives? 2. If it is with djmount, how do I tell djmount that the second drive only has 6 heads? 3. If it is MountList I should use, what are all the magic numbers? (specifically the unit number and cylinder numbers (i.e. do I start with 0 since the partitioning is done on the MS-DOS side, or do I use the actual physical cylinder numbers))? 4. (This is more out of curiosity and to show my lack of understanding of all the goings on with the bridgeboard): why doesn't Janus just use mount and format like the other disk drivers (especially since one can mention the driver to be used in the MountList)? Again, all insights or comments are appreciated. Now on to other comments prompted by my initial posting... Andy Finkel (andy@cbmvax.UUCP) writes: >AmigaDOS V1.2 only supports about 54meg per partition in V1.2. >This limit is removed under 1.3, using the FastFileSystem. Thanks for the tip. This size limit (along with whatever the minimum is) should be mentioned in the user's guide for the bridgeboard. Also, I seem to remember someone saying that Janus wouldn't support the FFS. Has that changed, or is it just currently true (i.e. before the official 1.3 release)? >Actually, it [dpformat] uses the exact same bad track method that is used >on the PC side for all of its XT hard disks. Good. That's what I thought I saw happening. I don't know if I'm on track here since I haven't been able to play much at this level yet, but I was guessing that Brian Rhodefer's (brianr@tekig4.TEK.COM) formatting problem might be that there were bad spots in addition to the ones on the manufacturer's flaw map and that Janus was pickier about it's disk surface than MS-DOS (at least during formatting). I would hope re-*lowlevel* formatting (not dpformat) with the errors given by DiskDoctor would take care of this. BTW, Peter da Silva (peter@sugar.uu.net) writes: >My assertion (which has been blasted again and again in this forum) is that >when the file system gets a bad block, it should mark it as unusable. I agree. But, let me do it manually. E.g. Requestor says, "Block ??? appears to be bad. Would you like to lock it out so the file system will no longer use it?" Me, "Yes." -- Randy Hendry randyh@ozvax.gwd.tek.com ...!tektronix!ozvax!randyh Tektronix, Inc. Graphic Terminals Division (503)685-3063w P.O. Box 1000, M.S. 63-523, Wilsonville, OR 97070 (503)639-7697h "In Oregon, the sun shines every day; in the summer, you can even see it."
andy@cbmvax.UUCP (Andy Finkel) (08/16/88)
In article <3101@tekig4.TEK.COM> brianr@tekig4.TEK.COM (Brian Rhodefer) writes: > > >>: Me(original posting). >: uunet!cbmvax!andy (replying). > unflagged: Me (riposting). > >Among the information I filtered out for brevity's sake (assumed it >was obvious) is that the reason I'm using a `Janus' harddrive for >the amiga is that it lets me have a hard drive on BOTH the 8088 and >68000 sides with only one physical device, (and only one cash outlay). >Accordingly, I have a large chunk of this hard drive mounted as a MS-DOS >filesystem (Bridgeboard's drive C:). Only 39M of the drive's formatted >capacity remained for the 68000 side, so that's how big JH0: was. What an interesting approach to asking for help...leaving out information for brevity's sake, assuming it was obvious. I'll have to try that approach in the future. Here's an offer...why don't you sent me *exactly* what you did, each step, and your *exact* hardware configuration, and your *exact* parameters into FDISK and ADISK, and we'll look at it and try to help you. Complete information goes a long way in cases like this. >More importantly, why isn't the >result of `dpformat' a usable disk drive? Obviously, because something is wrong. From the information you've given me so far, all I *can* do is play guessing games, based on the information you've given, and try to help. >P.S. >Contraction of "you are" == "you're". (for future reference) >Contraction of "it is" == "it's". What ? I'm trying to help a person who posts a P.S. like that ? I'm trying to help one of those people who assume that, if a typing mistake has been made, that the person on the other end doesn't know the correct form ? Hmmm...trying to help usually turns out to be a complete waste of time in these cases. But, hey, sent the info email. (Usually a request for email makes them stop sending those 'oh so clever' spelling/grammer corrections they make in an effort to show superiority over anyone foolish enough to actually respond to the request for assistance) -- andy finkel {uunet|rutgers|amiga}!cbmvax!andy Commodore-Amiga, Inc. "If we can't fix it, it ain't broke." Any expressed opinions are mine; but feel free to share. I disclaim all responsibilities, all shapes, all sizes, all colors.
andy@cbmvax.UUCP (Andy Finkel) (08/17/88)
In article <896@ozvax.GWD.TEK.COM> randyh@ozvax.UUCP (Randy Hendry) writes: >The saga continues... >onward... (i.e., now I'm back to the original problem:) Good. We're making progress :-) Let's see if we can continue to do so... >And did so for every track (cylinder) of the disk (until I stopped it that is). >There is a good reason for this: There are no heads (surfaces) 6-10 on this >disk! The disk has 6 heads (0-5). Now non-coincidentally :-( the disk on >which partition jh0: resides *does* have 11 heads (0-10). It appears >that djmount (or dpformat) is using the head information for the first drive >for all drives. I think you've got something here. DJMOUNT is asking the PC side for the configuration of the hard disk partitions, then constructing AmigaDOS device nodes. It sounds like its getting the environment vector wrong, however. If you know the partition information, you should be able to use the standard AmigaDOS MOUNT command, by entering this partition information into your MountList. >Seeing this error made me think that I might be able to tell djmount manually >about the number of heads by using MountList. I tried a couple of >combinations, but had no luck. Djmount doesn't seem to use MountList -- >is this true? Correct...DJMOUNT is statting the PC side... >3. If it is MountList I should use, what are all the magic numbers? > (specifically the unit number and cylinder numbers (i.e. > do I start with 0 since the partitioning is done on the MS-DOS > side, or do I use the actual physical cylinder numbers))? OK...here's where I slip into theory...(I haven't tried it myself) Since I know what DJMOUNT and MOUNT are trying to do it should be sound theory :-). I've been told that this procedure works...but *haven't tried it* *You have been warned* Unit 0 is JH0:, Unit 1 is JH1: Use the actual physical cylinder numbers. Put in # heads, surfaces, etc, to match your drive. Make sure you get it right. Then Mount it, rather than DJMOUNT it. (WARNING: If you get the parameters wrong, you'll write somewhere besides where you intended on your drive...) > >4. (This is more out of curiosity and to show my lack of understanding > of all the goings on with the bridgeboard): why doesn't Janus just > use mount and format like the other disk drivers (especially since > one can mention the driver to be used in the MountList)? Because the jdisk author wanted it to be automatic, rather than use the MountList. A good ambition, actually. >Also, I seem to remember someone saying that Janus wouldn't support >the FFS. Has that changed, or is it just currently true (i.e. before >the official 1.3 release)? If the above method of using the MountList works for you (as several developers have told us it does) you can use the FastFileSystem on the Janus disk. The reason we say its not currently supported is that DJMOUNT doesn't know the correct magic. And since moving FDISK/ADISK parameters into your Amiga side mountlist by hand is not the desired mode of operation... In this case, you would use FORMAT, rather than DPFORMAT. -- andy finkel {uunet|rutgers|amiga}!cbmvax!andy Commodore-Amiga, Inc. "If we can't fix it, it ain't broke." Any expressed opinions are mine; but feel free to share. I disclaim all responsibilities, all shapes, all sizes, all colors.