carroll@s.cs.uiuc.edu (02/23/89)
Sorry to bother you guys, but this has been ticking me off for a couple weeks
now, and I'm not any closer to figuring it out.
Last September I bought a Zeos '386 16MHZ Tower system with a 65M ST277R
drive, Adaptec 2372 RLL Controller, 1 M memory, 64K cache, 1.2 5.25" floppy,
and 1.44M 3.25" drive.
I stuck a Micropolis 1335 (71M MFM, 109 RLL) in, even though it wasn't RLL
rated, and it worked. Life was rosy. About 2 weeks ago, I got SpeedStor,
and I switched the machine from the Adaptec drivers to SpeedStor, on the
promise of more disk throughput.
To enable SpeedStor, I had to disable the BIOS on the Adaptec card. I also
re-organized the hard disk layout, making about 2 Meg on the ST277R and
all of the Micropolis 1 SpanDrive (the trick, ya see, is to run Norton
SD to move all the directories onto the ST277R, with the files on the
Micropolis, to reduce seek times). I was happy, since Norton SI reported
a disk index of 6.8, instead of the previous 3.5, and it was quite
noticeably faster.
Well, shortly after this, the system started hanging. As far as I can tell,
it hangs on disk operations, by DOS thinking the operation is pending, and
the controller not having heard it. It happens most frequently on executables
I'm trying to run, but it also happens while running PCTeX on long files.
It *never* happens while running Norton Utilities. I ran DT /f 6 times in a
row with 70 Meg of files, with no problems, and then the next executable locked.
The last access doesn't seem to be drive specific, in fact I've gotten it to
fail on the floppies a couple times.
Originally I could make it stop by formatting /s the boot partition,
but that doesn't cure it anymore.
I tried tracing the DOS vectors to look for viruses, but things seem to
be OK. I called Zeos, but they didn't know anything about SpeedStor. I
tried calling Adaptec, but I haven't been able to get through. I called
SpeedStor, and they indicated that I should enable Bus Wait States.
I thought I had, but the SpeedStor Tech claimed the manual was wrong.
I've tried with the jumper on and off, and I can't tell any difference,
except the disk is slightly slower with it on (which means the manual is
wrong). I've checked the cables, taken the card out and put it back in, etc.
The most *irritating* part of the whole mess is that it comes and goes with
the wind. It will get into a state where I can't even run one executable,
and then it will recover and run for a day or so with no problems. I have
no idea what causes it to fail or recover. Has anyone else had any problems
like this? Thanks!
SpeedStor - 5.13b AMI Bios 1/33/88 K
Alan M. Carroll "And then you say,
carroll@s.cs.uiuc.edu We have the Moon, so now the Stars..." - YES
CS Grad / U of Ill @ Urbana ...{ucbvax,pur-ee,convex}!s.cs.uiuc.edu!carrollcarroll@s.cs.uiuc.edu (03/05/89)
/* Written 8:28 am Feb 24, 1989 by davis@clocs.cs.unc.edu in s.cs.uiuc.edu:comp.sys.ibm.pc */ >A recent posting made me wonder why it was necessary to use third >party disk drivers in many cases. (Speedstore, Disk Manager) >the same time?) > >My real question is why bother with Speedstor. (...) >As near as I can tell, the only thing you buy by using Speedstor >drivers is to have a single large partition and your files can be >64Meg long instead of 32Meg. You may trade performance (I have heard >reports that speedstor slows stuff down somewhat) and compatibility. > No, actually SpeedStor greatly improved the disk throughput, and that's the main reason I switched. >A similar issue comes up with RLL drives. Why let the controller fool >the DOS about the number of sectors and cylinders. Can't DOS (after 3 >something) handle any combinations of sectors and cylinders (as long >as cylinders is ss than 1024). As far as I know, SpeedStor doesn't fool DOS about the number of cylinders and sectors.