Dead Cheetah: She was almost 7;-(
- Thread starter Santilli
- Start date
blakerwry
Storage? I am Storage!
heh, I still have some good quantum drives that have passed their 10th birthday... I haven't kept anything under a few hundred MB... so my pre-IDE drives are gone even though most still worked.
Dïscfärm
Learning Storage Performance
Santilli said:
Low level formatting* as in hanging the Cheetah off of an Adaptec AHA-2940UW, with the Cheetah being the only drive on the terminated cable, and with its SCSI ID jumper set to something other than "7", and having the Adaptec SCSI Setup set to scan all SCSI targets on bootup? I'm assuming the drive is spinning up and not convulsing (clicking, burping, farting, etc).
* Control-A Menu, SCSI Utilities, Format Drive... (PC version ofthe Adaptec host adaptor).
Santilli
Hairy Aussie
- Joined
- Jan 27, 2002
- Messages
- 5,285
NO.
I tried formatting about 4 times, using different scsi id's,and the bios in the card wouldn't find the drive in one utility, and, gave warning signs.
I still tried to reformat, as you indicated, but everytime I had an error, can't complete sort of message.
With the indicator light, that shows disk activity, I can usually tell if the disk is doing anything, and, in this case, it was no.
Niether the card bios, nor 2000 recognized the disk.
What kicked it was putting a very old Barracuda in the same spot, with the same scsi id, and having it recognized, and, reformatted by windoz.
How's Splash??
gs
I tried formatting about 4 times, using different scsi id's,and the bios in the card wouldn't find the drive in one utility, and, gave warning signs.
I still tried to reformat, as you indicated, but everytime I had an error, can't complete sort of message.
With the indicator light, that shows disk activity, I can usually tell if the disk is doing anything, and, in this case, it was no.
Niether the card bios, nor 2000 recognized the disk.
What kicked it was putting a very old Barracuda in the same spot, with the same scsi id, and having it recognized, and, reformatted by windoz.
How's Splash??
gs
Santilli
Hairy Aussie
- Joined
- Jan 27, 2002
- Messages
- 5,285
OK:
Al Capone's servant will read more carefully. I'll give a try formatting it on it's own, terminated, in this case, LVD cable, and see what happens.
Or, just terminate the drive, and see if it works on a GD UW cable.
I've got a bunch of cables around, I'll hang one off the internal, disconnect the raid box, and give it a shot.
gs
Al Capone's servant will read more carefully. I'll give a try formatting it on it's own, terminated, in this case, LVD cable, and see what happens.
Or, just terminate the drive, and see if it works on a GD UW cable.
I've got a bunch of cables around, I'll hang one off the internal, disconnect the raid box, and give it a shot.
gs
GIANT
Learning Storage Performance
Santilli said:
Well, if it won't pass the Adaptec test I outlined above, then there's definitely no hope.
Having the SCSI host adaptor send the low-level format command to the SCSI hard drive controller is pretty much the ultimate test for basic SCSI hard drive functionality.
If you have the drive out of the chassis and loose -- hanging off the SCSI cable (and powered, obviously) -- you might want to try pushing on the drive's integrated circuits or tapping on them with the handle of a light screwdriver. Sometimes solder joints go bad or connections open up over a long period of time and usage. I've revived a few drives doing this.
I've even encountered a couple of failed SCSI drives that would suddenly start working with a simple SCSI ID jumper change from "0" and few (more) that would only work if they were *not* set to "Autospin."
Sharkbitten once, twice shy. Otherwise, currently migrating to the mullet-rich waters of Key Largo.
Santilli
Hairy Aussie
- Joined
- Jan 27, 2002
- Messages
- 5,285
OK:
Pulled the case open, made it scsi 0, since it didn't work at 6, connected on the internal channel, with one of those nice, shiny, GD UW cables, jumped it to terminate on, and we are trying a reformat.
It's taking a long time, but the good news is, it hasn't quit yet.
Al Capone is pissed off, locked in the bathroom, since an open case with flashing lights and fans is a great place to poke little paws.
gs
PS If it fails this time, I'll give the autospin off jumper a try...
Thanks for the help
I hate letting old Cheetahs die without trying
Pulled the case open, made it scsi 0, since it didn't work at 6, connected on the internal channel, with one of those nice, shiny, GD UW cables, jumped it to terminate on, and we are trying a reformat.
It's taking a long time, but the good news is, it hasn't quit yet.
Al Capone is pissed off, locked in the bathroom, since an open case with flashing lights and fans is a great place to poke little paws.
gs
PS If it fails this time, I'll give the autospin off jumper a try...
Thanks for the help
I hate letting old Cheetahs die without trying
Explorer
Learning Storage Performance
The few "Auto-Spin" (a.k.a. -- "Auto-Start") jumper problems I've encountered in the past were all where the drive would NOT spin up at all until the Auto-Spin jumper was removed from the drive, followed with setting the SCSI host bus adaptor BIOS to "Send Start Command" to that SCSI target ID. In other words, hardware Auto-Spin quit working altogether, but a SCSI command fro the SCSI host bus adaptor over the SCSI bus would work!
In working with several hundred SCSI hard drives over the years (since mid 1980s, in fact) some other SCSI strangeness that I've ran into was an early IBM drive that suddenly one day became SCSI ID "3" no matter what I did to its jumpers. Neither of the two different sets of ID pins on the drive were physically shorted, but obviously a short existed somewhere -- in this case, somewhere in the hard drive's SCSI controller circuitry. The drive, otherwise, worked perfectly for many years -- but *only* as SCSI ID "3." I recall someone else telling me a while back about a SCSI drive that was stuck at SCSI ID "0" no matter what they did with the ID jumpers.
A more common problem that I've ran into on older SE (non-LVD) SCSI drives was poor or bad onboard drive termination. On long and/or multi-drive SE SCSI busses it's generally well advised that you go with an external terminator.
Not so much of a drive problem, but a problem that can really get one rather angry once you discover it, is BAD JUMPERS; they can look perfectly good, but they're most definitely bad. I've ran into bad jumpers maybe a dozen times. So, I'm always pretty cognizant of the possibility of bad jumpers when I encounter "SCSI Strangeness."
Just thinking about SCSI problems always makes me recall the time -- in 1991 or 1992 -- where I had a external Toshiba 30 MB SCSI drive with fairly important data on it to die on me (would not start up). I worked and worked with it for a couple of hours to no avail. I had pretty much discerned that the motor was locked up or had become weak, or possibly that the heads had stuck to the platters, because I could barely hear it trying to spin each time I applied power but was being held back. The drive always ran cool in its external case, so I never really believed that the heads were stuck to the platters (from excessive heat).
Finally, after a couple of days of occasionally tinkering with this Toshiba SCSI hard drive and getting nowhere, I decided that I would need to take some drastic manoeuvres to "free up the motor" by hitting the top of the drive mechanism with a medium-light-weight ball peen hammer with the drive laying flat on a carpeted concrete floor. At the same time I powered it on, I would give it one dead hit squarely in the middle of the top of the drive on-axis with the spindle. I did this 7 or 8 times increasing slightly the strength of each dead blow. There was still no change. Then, I decided to just deliver a single serious hard blow in the same spot on the spindle bearing. WHAM! Nothing. I figured that the drive would've probably cracked open from such a hard (and loud) hit. Then I hit it 3 times in swift succession just as hard as I did before. Still nothing.
I stared at the drive for about 10 seconds -- thinking. I walked over to my drawer and put on a thick canvas+leather construction glove to protect my right hand just in case the wooden handle should break, turned the power on, and immediately swung the ball peen hammer with a high-speed savage hit -- complete with my feet coming off the ground. Suddenly, the drive *slowly* began to spin! I just let it sit there and run for about an hour before turning it off and hooking it back up to the Macintosh's SCSI chain.
I thought that the casing would have shattered from such a hit, but it didn't. The ball peen hammer wasn't very heavy and the handle of just average length, so that save the drive. Incredibly, that stupid Toshiba SCSI drive was used for about 4 more years (non-valuable data).
Fushigi
Storage Is My Life
It's call drive stiction. It's a well-known trait of older high-speed drives that ran 24x7 for ages but suddenly fail to power up after being powered down. They heads are so warm that they actually melt/stick to the platter. This causes it to get stuck.
The other 'fix' is to pick up the drive and drop it. Non-powered, of course.
BTW, the hammer technique probably would have worked with less force by striking the side of the drive. For some reason sideways motion was a little more effective at breaking the stiction.
AS/400 AFAIK mainframes had this issue in the late 80s/early 90s. IBM somehow managed to deliver a fix via the drive's microcode. Dunno if they bumped the startup amperage, delayed shutdown somehow to cool the heads, or what. But it worked.
The other 'fix' is to pick up the drive and drop it. Non-powered, of course.
BTW, the hammer technique probably would have worked with less force by striking the side of the drive. For some reason sideways motion was a little more effective at breaking the stiction.
AS/400 AFAIK mainframes had this issue in the late 80s/early 90s. IBM somehow managed to deliver a fix via the drive's microcode. Dunno if they bumped the startup amperage, delayed shutdown somehow to cool the heads, or what. But it worked.
GIANT
Learning Storage Performance
Fushigi said:
I was aware of drive stiction at the time. That Toshiba drive had a different problem. I had stiction problems with a particular Maxtor full-height (5.25-inch) drive... er... something like maybe a Model 8270S, if I recall correctly. This particular boat anchor Maxtor model was one of the faster drives available at the time for recording real-time data because it had no thermal recalibration routine in its firmware. I developed a bit of a trick to avoid its stiction problem by simply allowing the drive to spin idly for about 20 ~ 30 minutes after a recording session, thus allowing its deep internals to (hopefully) all cool down to normal operating temperatures before shutting it off. This included running a "Park" command via a low-level utility before shutdown. It worked.
The Toshiba drive mentioned earlier had a problem that you can get with hard drives that simply have a lot of run time on them. The problem was dust buildup getting into the spindle bearings.
First of all, this was at a time when less-expensive hard drives typically had "spluttered" platter media. Spluttered platters are simply platters that have the recoding emulsion sprayed onto the disc substrate much like tape or floppy disc media have their recording emulsion sprayed onto their polyethylene or mylar substrate. "Better" hard drives used platters with an electro-plated magnetic recording surface. (Even the cheapest hard drives today use plated platters.)
Anyway, if there was enough dust build up inside these older hard drives, some of this dust could penetrate the seals in the spindle bearings and drag down or seize the bearings.
Oh, I did that -- on its side top, bottom, ends there on the desktop. I was thinking that there was a *slight* possibility that the heads were stuck, but not much. I didn't get into a lot of details in my earlier post, but the motor definitely seemed like it was pushing the spindle a little bit each time I hit the power -- maybe 1/32 of a turn each time. So, after playing around with for a while at the desk (as I did other things), I took it over to a place in the room that had a carpeted floor (this was in a room with a bunch of raised floor) and started banging away. Finally, I relocated a few feet over to a place where the stairs came up to the carpeted concrete floor, situated the drive at waste level, turned the power on to the external SCSI housing, put on my glove, and gave it a brutal gorilla hit. At that pint, I didn't much care if the drive housing shattered. It didn't, and the damned thing began to slowly spin up.
I can recall more than a few "bumping up the startup current" fixes that some drive manufacturers did to hard drives. This has been done with a firmware update for the onboard programmable logic that controls the spindle motor and sometimes with resistor value changes on the PCB. It was to get hard drives spinning that wouldn't, due to the internal dust buildup problem or from overly stiff bearings.