Debugging the Power Mac 9600 G4

[GorfTheChosen]

You seem to be doing more diagnostics and burn-in testing than even the engineers did at Apple years ago 

If so, that ain't saying much ... 

[GorfTheChosen]

Was testing out Retro and the Sony DAT on the Gigabit machine (under OS X 10.4.11) with the Adaptec 2906 card I stuck in it ... all sorts of errors on compare after back up ... to the point that the back up/compare would not complete in any kind of a reasonable time.

Made sure I had the latest drivers from Adaptec installed ... still no go.

Tried swapping out the SCSI cable for different one ... nada.

Finally moved the drive over to the 9600's external SCSI bus and did a couple of back ups (> 1 gig) under OS 9 ... no errors.

Later ran across a post somewhere that said the drivers for the 2906 (and possibly the 2930, since they use the same kext) are not compatible with anything later than 10.3.9.  Had previously run into some posts (somewhere) that said OS 10.4.9 and/or 10.4.10 had broken the drivers for some of Adaptec cards.

Still need to retest it on the 9600 using the cable that I was using on the Gigabit ... just to eliminate the cable as a possibility.

[ssp3]

I'd suggest you check the web archive more often before conducting various old PCI card experiments 

http://web.archive.org/web/20120415194823/http://www.adaptec.com/en-us/support/scsi/2900/ava-2906/

[GorfTheChosen]

I'd suggest you check the web archive more often before conducting various old PCI card experiments 

http://web.archive.org/web/20120415194823/http://www.adaptec.com/en-us/support/scsi/2900/ava-2906/

Yeah ... been there already, done that before embarking ...

What would you have me "check" exactly ?

There's no definitive statement of incompatibility there with later operating systems that I saw.

And, if I'm not mistaken, the "supported operating systems" roughly corresponds to Adaptec product development on the Mac ... to the point in time where they threw in the towel and discontinued developing drivers for Mac completely.

[GorfTheChosen]

Some progress to report on the 9600:

Managed to get it to boot Mac OS X 10.4.11 using XPostFacto4.

Had to do the install of 10.4 on an IDE drive (Maxtor 60GB) in a mobile rack carrier connected to the Quicksilver G4 800 Dual via FW. Went ahead and used Software Update to apply all the Apple OS and Security updates while still attached to the Quicksilver.

Then moved the mobile rack to the 9600 and inserted it in the receiver which was attached to IDE Channel 1 on an ACARD AEC-6260M PCI host adapter. Used XPF4 to apply the necessary components there and successfully boot into 10.4.11.

Parameters were 1. to make sure the install was done to the first partition on the drive, 2. the partition was less than 8GB in size, 3. the L2 Cache checkbox in XPF4 was unchecked (L2 Cache NOT enabled during boot) and 4. the speed during boot was set to 8 ... although I've since slowed it down even further to 10.

After doing that and getting it to boot successfully, I partitioned the OWC Mercury 3G SSD - which is connected to the ACARD AEC6260M IDE Channel 2 with a red Startech IDE to SATA adapter - to comply with the "less than 8GB partition rule" and used Carbon Copy Cloner to clone the boot partition off the Maxtor drive onto the first partition on the SSD.

That booted successfully as well ... so booting off IDE Channel 2 certainly seems like it works (I heard heard something to the contrary)

I then used CCC to clone the OS X boot partition on the SSD onto its 5th partition - which is almost 40GB into the SSD - to see if that would boot. It did ... so apparently using IDE host adapter resolves the "must be in the first 8GB rule" ...

Overall, it seems pretty stable so far ... obviously the performance is slower compared to the other machines here and the video redraw can be a little laggy at times ... but seems respectable given the capabilities of the machine.

Am having an issue with file copy performance across the network via the Gigabit NIC ... but I suspect that might be a driver related issue (or possibly a bad Cat 5e cable)

More to report later.