Target Disk mode on unsupported MDD

[macarone]

>Wow, never heard that one, let's clarify, FW Disk mode on an MDD 2003 FW400 Single or Dual will work 100% of >the time, I literally used to do that to prep dozens of machines back in the day.

It's been corroborated that SATA drives connected via either a PCI card OR an ATA to SATA adaptor are not only recognized, but will boot an MDD.

If you only have drives connected via a PCI card, and NO drive connected to the ATA 100 port, the MDD will shut down shortly after you try to enter TDM.

It was often reported that, a FW400 MDD would fail to mount an external FW drive, and the solution was to disconnect the power cable, etc. to reset FW in the MDD.

This drive enclosure looks nice:

     https://oyendigital.com/hard-drives/store/CB3R2-SL.html

     https://www.dpreview.com/forums/thread/4085999

It has the advantages that it accepts any height 2.5" laptop drives, of any size, including 2 TB, and in addition to RAID 0 and RAID 1 it also does JBOD and BIG. It runs from FW computer bus  power without the power adaptor.

BUT: An MDD will recognize it as a bootable FW drive in Startup Disk OS 9 or OS X, however, it will in fact NOT boot.

WORSE, and I know this sounds flaky, but after trying to boot Leopard from this device, you will find that the Leopard partition has been corrupted. Partitions containing only OS 9 or Tiger remain intact.

This case only works right if it is plugged in AFTER the computer has already booted.

Funnier still, some on Windows PCs, have also reported not being able to boot from it using USB, FW or eSATA. The company denies all problems.

[GaryN]

This is turning my brain into a pile of mush. We now have two more conflicting opinions.
Hyram "agrees" that only drives on the ATA100 bus will be seen in TDM and Macarone says they "work 100% of the time" but doesn't specify the bus.
I can't be certain that he means no bus issues or he just never had a problem
This is what's been happening to me the entire time I've been trying to figure this out - just enough non-specific uncertainty to prevent a definite answer.

Fury deBongo: My friend! (and person who actually has a machine out in "test mode"
Can you confirm this one way or the other: Can your MDD be "seen" in TDM with the boot drive on the ATA66 bus?

My head hurts…

[FdB]

Fury deBongo: Can you confirm this one way or the other:
Can your MDD be "seen" in TDM with the boot drive on the ATA66 bus?

My head hurts…

Give me an hour or two (maybe less)
and I'll get down there and check that specific question out.

And yes, my head also hurts.

[DieHard]

Wow, never heard that one, let's clarify, FW Disk mode on an MDD 2003 FW400 Single or Dual will work 100% of >the time, I literally used to do that to prep dozens of machines back in the day.

Hey Gary, I think he was simple quoting me...

And yes, as I explained, the FW target mode (sending commands to the hardware programmed via the firmware) is limited to disk I/O connected to the logic board, which makes sense. Yes "back in the day" we did not use SATA PCI cards, we used PATA/IDE mechanical drives.

These days SATA SSD drives connected vis SATA/IDE adapters (where we are still connecting to the MB IDE channels) are fine and all legacy IDE drives should work 100% of the time. 

But "back in the day" FW 800 MDDs were NOT running OS 9 and my company was making OS 9 DAWS so I cannot comment on them.  But the FW400 2003 model was rock solid as far as the target mode.

Also, as we surmised SCSI drives connected to PCI cards in MDD is a source of confusion since many of us veterans remember when there was "SCSI disk mode". which was around prior to FW disk mode. Powerbooks could become external SCSI drives and be connected to the chain to the SCSI port of a beige G3 and such, however SCSI cards in G4 towers render the drives invisible when using the more modern "FW Disk target mode"; we conclude also that any Mac with FireWire added by a PCI or CardBus card, are not supported.

Kinda beating dead a dead horse... just stating the obvious to most, but maybe not to all.  I will say in closing that "FW target mode" when used as intended is NOT flaky and is a life saver.  I did (2) data recoveries today on an ibook g4 and a Mac Pro 2007, both of which I did NOT want to open. They sat on top of a Mac Pro 3,1 running El Cap, and all the data was recovered from 2 machines that no longer boot. 

[macarone]

Kinda beating dead a dead horse... just stating the obvious to most, but maybe not to all.  I will say in closing that "FW target mode" when used as intended is NOT flaky and is a life saver.  I did (2) data recoveries today on an ibook g4 and a Mac Pro 2007, both of which I did NOT want to open. They sat on top of a Mac Pro 3,1 running El Cap, and all the data was recovered from 2 machines that no longer boot.

And using TDM in such a situation often can make the computer that is not bootable run again by running repair utilities on the hard drive.

I was in no way knocking TDM, only cautioning that it only works on drives connected directly or via adaptor to an ATA port. I also tried to explain that a computer in TDM with the FW icon bouncing around the display may not show up on the desktop of an MDD with a frozen FW system, where NO FW external drives are mounted. The fix is to leave the computer in TDM as it is, but to shut down the OTHER computer, remove the power cord, hold the power button down for 5 secs, and then plug the power cord back in and reboot.

[DieHard]

No argument here; We need everyone's experience and opinions

I am just glad it exists, nothing like it in the PC world and it has saved my bacon.  Dis-assembling iBooks and removing the entire tops off some macbooks is a drag just to get to the hard drive.  I have recovered data from bad hard drives in iMacs without having to remove the screens, so I am very grateful to apple for at least this making this feature.  Every hard drive dies a different death. In some iMacs the drive literally locks up the unit from Booting to an external hard drive, yet FW mode till works, go figure

[FdB]

Can your MDD be "seen" in TDM with the boot drive on the ATA66 bus?

In this particular instance,
on this particular MDD (1GHz FW800)
the answer is a resounding NO.

Will not boot into TDM with only a HD on the ATA 66 bus.
Added an OS 9.2.2 pre-loaded HD to the ATA 100 bus,
while also keeping the OS 10.4 HD on the ATA66 and it
then boots into TDM. Haven't checked whether or not
"other machines" can "see it" yet, but without the 9.2.2
HD on the ATA100... it's-a-no-go.

AND, as I just happened to grab a HD with OS 9.2.2
already on it... the fact that it then boots into TDM
(whereas the 1.42 GHz MDD would not) the
"differences" between the two MDDs prompts
further, closer examination?

Later.... that same evening...

AND, now… booting into TDM, the 1GHz FW800 IS
recognized and mounted upon the desktop of a 1.5
GHz PB 5,5 G4. Inversely, the FW800 does the same
in reverse, except the PB is unseen / unmounted when
booting the MDD in 9.2.2. BUT, it does see the
PowerBook when the MDD is booted in OS 10.4.
(MDD's 10.4 HD, still on that ATA66 bus.)

Whatever is going on with that FB person, Gary…
it sounds related to whatever is happening here
with my (first tested) 1.42GHz MDD FW800.

The 1GHz FW800, seems to work just fine in TDM.
(Running same Open Firmware version as the 1.42GHz.)

This “patient” may never attempt the Open Firmware
FW400 “regression therapy” for his ailing 1.42 GHz
FW800 MDD TDM... until the snows fly again.

And doesn’t the old “scuzzy disk mode” still have that
“unclean” ring to it? It did work very well at the time.

I gotta get me another horse!
But, I will swap the drives from the 1GHz into the 1.42GHz
eventually and see if that makes any freakin' difference on
the 1.42GHz.

If it does, I'll report. Otherwise... "crickets".

[GaryN]

I have a sneakin' feelin' that's exactly what's happening. he's got one drive in the MDD and maybe it's in front - so no go.
He gets a little slack because it's his first MDD, and we all know it takes some hard-learned lesson time to get a grip on the insanity idiosyncrasies of this beloved model.
I'll tell him to move it to the ATA100 and see what happens. If that fails, he'll have to decide if he wants to go to firmware regression-land.
I hoping that this is all it takes though…

You're the best Fury! Many thanks!!

[FdB]

You know that in this case, TDM stands for:
Taxes-Death-Macintosh?
(Or This Damned Mac.)

Begin-pulling-hair-out-now.
(I am now bald.)

Moved HDs from 1GHz FW800 back to the 1.42GHz.
(10.4 on OWC SSD to ATA66 and conventional HD
with OS 9.2.2 to ATA100.)

Removed battery checked voltage ≥ good. Removed and
replaced with new one anyway. (Same voltage reading).

Booted, holding Option+Command+P+R for the 3 chimes
and it booted from the OWC SSD. Set clock, time zone, etc.
Rebooted, Option-key and selected OS 9.2.2 drive on ATA100
as boot drive. Booted and then shut it down.

Started up again with T-key and it booted right into TDM!
FW800 to 400, 400 / 400 and 800 / 800 all provided mounting
of the OS 9.2.2 drive on the G4 PowerBook running OS 10.4.

Even connected to a Mac mini (FW400) as the G4 PowerBook
did not have OS 9.2.2 present and the mini both recognized
and mounted the conventional OS 9.2.2 HD in TDM, under
both OS 9 and OS X boots of the mini.

However, “No see” or never-saw the OS 10.4 HD.

May-be that macarone's suggestion of unplugging overnight
has some bearing here? Hell, I even did a PMU reset before.

However, it is highly probable that in the first HD config of the
1.42GHz, that OS 9 and X occupied the same HD on ATA 100…
on two separate partitions… and now, under this final dual-boot,
separate HD config on the 2 separate ATA buses… it just freakin’
(as if miraculously) works. (Why no angel-with-halo emoji?)

Thusly:

(1). Perform a fresh “solo” install of the OS 9.2.2 on HD by itself.
        [Placed last (& possibly alone) on the ATA100 ribbon cable.]
(2). Place other, secondary drive (for now at least) on the ATA 66.
        (Format for bare “storage” drive, or install a version of OSX.)
(3). Check & maybe even change battery, regardless of voltage.
(4). Rebuild / ZAP PRAM at startup. Set clock and Time Zone.
(5). Shut down, or restart…
      (a). After choosing TDM boot option in Sys Prefs of OS 10.4.
      (b). OR, by holding down the T-key... if cold booting.

If the above works, then one might then consider moving that
ATA66 HD to the middle position of the ATA100 ribbon cable?
(Unsure of an OWC SSD with ADP-06 bridge, in that position.)
I've been messin' with a Quicksilver & Master/Slave/CS stuff.

OR… simply install a fresh battery, rebuild/ZAP and try it again!
                But most definitely, OS 9.2.2 HD on the ATA100 bus.

And no thanks necessary, Gary. If not for you and that Kaffee
LAH-TAAY-swilling DieHard... such bloviated, MDD exposés as
this would not be here... as I would have simply given up on
my own, first MDD resurrection. Thank... you.

[GaryN]

And no thanks necessary, Gary. If not for you and that Kaffee
LAH-TAAY-swilling DieHard...

Thanks agin' F.

[DieHard]

Hey Gary,

Fury got me to thinking, ask your guy if ALL hard drives are jumped as "Cable Select" as they should be on all MDDs, Maybe the Target mode expects this.  Maybe it's more of the jumper position than which IDE controller. also, ask if the IDE cable if the factory Apple cable that has a "hole" in one of the lanes, that also designates cable select, maybe the IDE cable has been replaced.

Now the plot thickens on the cable select "hole", I was told by a Seagate rep (back in the day) that the hole was definitely used for cable select feature and now I found this:

Ultra-ATA 80-conductor cables were introduced to reliably support data transfer rates faster than ATA 2. The additional 40 conductors are grounds, designed to reduce/eliminate noise and crosstalk between conductors. The presence of the hole in a single conductor near the host controller's connector on an 80-conductor ribbon cable doesn't have anything to do with determination of Cable Select. Pin/conductor 34 on an 80-conductor ribbon cable serves two purposes, one of which is to provide the host controller a means of identifying the type of cable being used (CBLID) - either an Ultra-ATA cable with 80 conductors or an older/slower ATA cable with 40 conductors. This is accomplished by connecting pin 34 to ground in the cable's host controller's connector. Beyond the host connector, that conductor must be open - which is the hole that you see in the cable. Between the device connectors, conductor 34 is used by the attached Devices 0 & 1 to communicate with each other. Device 0 sends an Execute Device Diagnostics command to Device 1, and within a pre-determined amount of time, Device 1 must send the PDIAG (passed diagnostics) response to Device 0. If Device 1 has sent a PDIAG signal, it supports ATA-3 or faster protocols. In essence, the purpose of the hole in the cable is to identify whether the cable and connected devices are capable of supporting Ultra-ATA data transfer speeds. Newer Ultra-ATA cables omit the hole, by not having an insulation-piercing tooth at pin 34 in the host connector, and shunting that pin to ground.

To enable the use of the Cable Select jumper setting on connected devices (and when supported by the host controller), the Cable Select (ground) signal is conveyed on conductor 28. The device connected to the end (black) connector detects the ground signal and configures itself as master. Because the insulation-piercing tooth at pin 28 on the middle (grey) connector is missing, the connected device doesn't detect the CS signal and configures itself as slave. Incidentally, reversing an Ultra-ATA ribbon cable shouldn't be done. If the host (blue) connector is attached to a drive and the master (black) connector is attached to the motherboard, the intended CBLID/PDIAG functions of the cable are negated, which will affect drive performance.

So the bold sections are interesting (the rest seems on par with my brain); AFAIK speed negotiation was via the controller and the HD PCB board, this guy says the hole is used for that... hmmm I wish I could half of all the great stuff I forgot

[GaryN]

Hey Gary,
Fury got me to thinking, ask your guy if ALL hard drives are jumped as "Cable Select" as they should be on all MDDs, Maybe the Target mode expects this.  Maybe it's more of the jumper position than which IDE controller. also, ask if the IDE cable if the factory Apple cable that has a "hole" in one of the lanes, that also designates cable select, maybe the IDE cable has been replaced.

He immediately blew up my guess that his boot drive wouldn't appear because it was on ATA 66. It was on the ATA100. Also, he's tried swapping out other drives from his other G4 front, back, upside-down etc. without success. He has NO trouble booting the damn thing however with any of them: 10.4, 9.2 in any position and the FW bus and functions are completely normal in every other respect. AARRRGH!

I have provided him with the links to the firmware downgrade topics along with a suggestion that maybe he should re-evaluate just how important having target mode on the thing really is to him before starting down THAT road. This has become one of those exercises in futility that pop up now and then. It probably IS a firmware issue since literally everything else has been looked at.

Thanks to all for your efforts - if anything illuminating develops, I'll let you know.

[IIO]

that would have been a great outcome if it would have been the jumper settings.

[GaryN]

that would have been a great outcome if it would have been the jumper settings.

Nah…too easy. Where would the fun be?

[refinery]

its actually always been a requirement of TDM that the connected drive is on "bus 0"... this is documented even on Apple:
https://support.apple.com/kb/TA25584?locale=en_US

i can clearly remember having to play drive shuffle back in the day when doing some data recovery for people to access all the disks over TDM. The other busses simply wont be seen.