#Cart drive hangs boot
25 messages in this thread
Rick,
Easy solution — leave a cart in the drive. I did the same when I had a
Syquest and do the same now with a CD-ROM.
-sja
Steve,
> Easy solution — leave a cart in the drive.
Isn't that really more of a "workaround" than a solution?
I'd rather not leave a cart in if I can avoid it. This machine has to stay up
24/7, and in the event of the watchdog cycling power to reset it, or a power
failure, it has to come back up unattended. Which means the cart would have to
stay in constantly. Since it's only used for backup it seems a shame to suffer
all that wear and tear on the spindle motor. Not to mention the extra stress
on the power supply and the extra heat in the box.
I've actually been leaving the cart in already. But I hope it's only for the
interim. I know this is doable, because I've seen it done.
Good to chat with you again after all this time.
Rick
Rick,
I don't know … when I had a Syquest a cart was in it at all times (one of a
dozen or so carts I had). Now days, I use scsi tape streamer, and it's always
got a tape in it. It's no different than a hard disk spinning, but, to each
his own.
Steve,
> Now days, I use scsi tape streamer, and it's always got a tape in it.
Yeah, but is it true that for a tape drive the motor isn't running unless
you're accessing the tape? Not exactly a comparable situation, if so.
> It's no different than a hard disk spinning.
<Shrug> To be brutally honest, Steve, you're probably right. I guess it's just
a bias I'm carrying based on how I've seen something work, and therefore think
this _should_ work. Plus maybe the perception of the thing more as a floppy
drive than a hard drive, since the medium is removable.
Dunno. Humor me? <Smile>
Rick
P.S. Streaming tape now, eh? May I ask why you went this way? When I was
looking for backup media, the way things priced out I had decided to buy a big
SCSI hard drive for backups. I could easily get a 3.5" one-gig drive,
removable mounting slides, and an external box to take the "plug-in" hard drive
for hundreds less than a more-or-less equivalent sized tape drive. Plus I
figured I'd get faster transfers and random access. (The only reason I got the
SyQuest, in fact, is that a friend had it sitting brand-new and unused on a
shelf, and only $350 for it with a cart.) Of course one tape drive and ten
tapes is much cheaper than ten hard drives, but still…
Was I not finding good prices on tape drives, or is there some reason to prefer
them that I missed? Wondering for future reference.
I've got a streamer because I'm running Unix on my 3000 and a streamer is
required to install the system.
-sja
Ah, well that answers THAT question. <Grin>
Rick
PMJI, but I just saw the thread….
Have you checked to see if the cart is the last drive (by number) on your
SCSI chain?
For a long time, I was wondering why my A3000 started to take 'forever' to
boot. I found out that HDToolBox sets a flag on the RDB of the last drive
in the chain, telling the OS to ignore disks above that address. If the
cartridge drive was connected the last time you updated the RDB of your
'next highest' drive, it will set the flag to telll the OS to look for
another drive. If that drive is a cartridge – w/o a disk inserted – it
will sit there for up to 10 seconds, waiting for some valid drive to show
up.
_IF_ your cartridge drive is the highest drive in your SCSI chain:
To avoid a delay during reboot, you have to turn off, or physically
disconnect the cartridge drive, then update the RDB of the 'last' hard
drive in your SCSI chain. (The program RDPrep actually shows the status of
this flag – otherwise I'd never have known about it.) If you do this, your
machine won't wait for a (non-existent) cartridge to come up to speed.
Erik Flom
> PMJI…
"Pardon???" But jumping in is what I _want_ you to do! <Grin>
Interesting tidbit of information; thanks for sharing it. Currently the
cartridge drive is #4, followed by two hard drives at #5 and #6. I'll remap
them and try your suggestion if nothing more suitable turns up. I'd _really_
like it to just ignore the drive if there's no cart in it or mount it if one's
loaded, but your approach is an excellent alternative, and it'll certainly
serve the purpose!
Thanks for the input.
Rick
Erik,
I just noticed something I missed when I composed my first reply to you. You
said the system "will sit there for up to 10 seconds, waiting…" I think I
could almost live with that! The problem I'm having is actually that it waits
_forever_ (well, at least several minutes) and never goes on. Didn't know if
you'd have any comment about this or not but thought I'd mention it.
Now I'm off to pull my A3000 apart and reassign the drives. Letcha know how it
works out.
Rick
Oh, I thought you had only a short delay (which is what I ran into when I first
installed a cartridge drive)….. Are you sure your termination, and terminator
power is correct? (I remember one external Syquest mechanism would prevent my
3000 from booting – by turning on the power to the external drive _after_
booting, I oculd then mount it using SCSIMounter……
Oh well, Good Luck! Erik Flom
Erik,
Well, I have done as you suggested, and the symptoms have changed.
Let me echo what I did to make sure I understood you. I remapped my drives so
the 60 meg HD is #4, the 212 meg HD is #5, and the cartridge drive is #6. I
left the cartridge drive off the SCSI cable, booted the system, went to
HDToolbox, and saved the changes to the hard drives. Then I powered down,
plugged in the cart drive, rebooted, and saved changes to the cart drive
_only_. If I got what you were saying, that should have left #5's RDB thinking
it was the last drive in the chain.
As I said, the symptoms changed. Before, you couldn't boot without a cart in
the drive. Now, from a power on the system hangs with the cart access light
lit steady; Kickstart doesn't even get loaded (I have one of the old developer
machines with Super Kickstart on the hard drive instead of in ROM.) But if I
then reboot with Control-Amiga-Amiga, it comes up normally.
Sadly, since this machine has to restart unattended, this won't quite get me
where I'm going. But it is better, and I thank you for that.
Any other suggestions would be welcomed.
Rick
OK, sounds like the problem now might be that it's not finding the right
SuperKickstart partition? I can't remember how that was set, but are you sure
that the system partition with the ROM image is the one it boots to? (I can't
remember whether it looks to the lower or higher ID#….)
I know I got a Syquest to work with my A3000 before I put an 040 in it, so I'm
sure you'll iron out this last bug.
(Of course, this is the one day in months that I trashed my message base, and
this is all being done online….) Erik Flom
Erik:
I'll respond to all three of your messages here to keep the thread from
expanding further. I know some of your earlier questions "self-answered" when
you read my later messages, but I'll hit them all anyway.
The problem isn't a delay, but a hard hang (I know you know this now). Relative
to your question about when it delays, it only hangs if there is no cart in the
drive on power-up. Otherwise it works fine.
Are my terminators okay? Dunno, maybe not. I left them in the SyQuest because
it's the last unit on the cable. Before that is the Conner, and I know it has
no terminators. The first drive on the cable, the Maxtor, I'm not sure about.
Maybe it still has its terminators. Do you think that would cause this sort of
problem?
I know what you mean about the power-off trick; I used to have access to a
system with an external CD-ROM. Either the power had to be off or a CD had to
be in it, or it'd hang the system on boot (just like this). But since this is
an internal drive (yes, it's an A3000) I can't kill power to it.
I checked with HDToolbox and the cart is marked non-bootable. Not sure the
search order would cause this symptom; if it was trying to load KS from the
cart drive then I would expect it to hang with a cart in the drive too. Anyway,
I guess that's been eliminated now that the SyQuest has been both unit 4 and
unit 6 (i.e., regardless of the search order, it's been in the right place at
one point or another).
Rick
Thanks for the detailed response…Got it all clear now.
So, you've got 3 drives, _all_ internal?
(Must be toasty warm in there! :^)
I'm at a loss to answer your question, though! :^(
I've been sitting here, but every hypothesis has been checked……
All I could recommend is to try using Microbotic's RDPrep (which used to
be available in their forum, when they had one…..), or some other RDB
util, to see exactly what flags are set and which filesystems you're
using.
I wish I still had a SuperKickstart machine (mainly to load 3.1, but
that's another story)! I have a feeling it has something to do with that,
but don't know how….
As for terminators, just make sure that Maxtor doesn't have any. (I can't
see as to how it would cause your problem, but better to check anyway.)
Maybe someone with a SuperKickstart machine might chime in with any ideas?
Good Luck!
Erik Flom
Erik: I got GP-Fax today. Haven't had a chance to try it yet. I"ll tell you if
it works. Thanks! Tom Hunt
Erik,
Well, four drives if you want to count DF0:. And I still have room for one
more, but so far I've been able to resist the temptation. 8) As for internal
temperature, I don't think the drives are generating any more heat than some of
the accelerator boards, though that's just a gut feeling.
I checked the Maxtor last night after I left my message. Sure enough, it still
had the terminators in it. Unfortunately, pulling them didn't make any
difference.
I will look at the RDB and see if there are any clues hiding in there, thanks.
Rick
Well, best of luck! I wish I could figure out what's causing it to hang?!?
At least _I_ seem to have _finally_ gotten the 40MHz GVP 040 working in
my
A3000!!!
Ever since I installed it, the 3000 SCSI would lock up when I ran more
than one application at a time (even little utils!) After much wrangling
with GVP, they finally admitted that I'd have to replace 'all my custom
chips' to get it to work right (!) After months of searching, I finally
got the third (and hopefully final) chip – the rev 4 SuperDMAC!
Thus far, it seems to be much more stable!
I can only hope you are as lucky – keep the faith! :^)
BTW, you were right about having to _unplug_ (power-down) the Syquest
before running HDToolBox – even w/o a cartridge, HDToolBox will set the
previous drive's LastDrive flag to look for the Syquest next boot
time……
Erik Flom
>I wish I could figure out what's causing it to hang?!?
Heh… you and me both. But at least it's non-fatal. As Steve has pointed
out, I can simply leave a cart in the drive until I find a solution. Not like
your problem, which made it difficult to do much that was useful, I'd imagine.
Don't be too hard on GVP regarding the custom chips. I was involved in an
Amiga project that used DMA. When the hardware was all finished and in beta
test, we began finding bugs that turned out to be in the A3000's custom chips.
In fact, the company I worked for at the time brought one of the bugs to C/A's
attention. They didn't even know it existed, yet it rendered DMA almost
useless in some circumstances. These bugs were fixed in the chips for the
A4000, but darned if we didn't turn up some new ones in _that_ machine!
The intent here is not to slam C/A; the Amiga is a complex machine and it's
difficult to get everything perfect on the first pass. Rather, the message is
that GVP saying "It's a problem in the Amiga" is probably not a matter of them
trying to find a scapegoat; they were quite possibly bitten just as we were.
Anyway, glad to hear it's all working properly now!
Rick
I guess I'm a little rough on GVP…..the original 28MHz 040 I had in my
A3000 worked flawlessly (even if it did create enough heat to fry eggs
:^), and the 40MHz accelerator worked fine in my 4000 for several months
before being transplanted to my 3000 – and, now that I've got all the
chips, it is once again a joy to use my 3000 w/o fear of a random lockup!
Best of luck…
Erik Flom
Richard,
I also have a Syquest (105) in my 3000.
If there is no cart in when the machine is booted, it takes about 30 seconds
for the harddrive to go ahead and boot. With a cart in the Syquest, it boots
from the HD at once.
Using HDToolBox, My system shows the Syquest as unit 0 and the Maxtor as unit
6, both with a LUN of 0. I presume this means that with no cart in the Unit 0
drive, the system waits until the SCSI interface times out before going to the
next possible drive on the bus.
If this is the case with your system, you could try changing the Unit number
jumpers on the two drives. Another possibilty would be one of the programs that
allow you to change the SCSI parameters, such as time out. You might be able to
"force" the system to look elsewhere for a bootable drive after a shorter wait.
Hope this helps,
Rob Hays
Rob; If one does not typically have a cart in the drive then it is usually
useful to move the boot drives SCSI ID down (ie: 6 to 5), move the SyQuest (or
any other removable) to a value greater than the fixed drive(s).
With no cart(s) installed, run HDToolBox (RDPrep or whatever) and set the "last
drive" flag.
The result will be that the system will not scan for the removable at all. To
use them, you run a utility such as scsimounter.
-bill
Bill,
> … move the SyQuest… to a value greater than the fixed drive(s).
>
> With no cart(s) installed, run HDToolBox… and set the "last drive" flag.
>
> The result… the system will not scan for the removable at all.
Sorry, did exactly that two nights ago at Erik From's suggestion. Didn't make
any difference. Also tried it with the SyQuest _unplugged_, figuring that if
it was connected it might still be found by HDToolbox somehow.
I'm going to peek at the RDB tonight and see exactly how the flags are set,
just in case.
Rick
Rob,
The problem I'm having is not a long timeout, it' _no_ timeout. Just to make
sure I waited five minutes just now. The system is just hung, period.
I've already tried the SyQuest as the lowest HD number and as the highest, and
have also set the "last drive" flag on the drive before it (or did if I
followed intructions correctly). Nothing has helped yet.
Rick