PDA

View Full Version : ATTO (1.6.4), OS 9.2.1, SoftRAID



TZ
09-09-2001, 11:52 AM
Freeze on boot when Finder/Desktop should show up.

Happens when even simple hfs+ on internal drive is used (Atlas 10k II formatted with 9.2.1 and pretty basic extensions, clean install). I was using mirrored but it seems problematic on the ATTO in my B&W.

I don't have trouble booting from external drive on the ATTO formatted with Drive Setup 2.0.7 and same extensions.

Rule out everything but ATTO, SoftRAID - and how they get along on MY system.
I have to zap PRAM and that doesn't work well. My plan today is to either wipe the RAID totally or just continue booting from external drive.

I had some trouble identical to this booting off a brand new 2G X15 Cheetah when it was internal and formatted with SoftRAID also.

It could be ATTO + B&W + SoftRAID + 9.2.1. And that is a "strong" could that for me, I'm looking to ATTO and others for confirmation.

Energy Saver also now triggers problems I didn't have before (but was never perfect either):

IF a disk drive goes to sleep (set to 60 minutes but writing doesn't require disk access), sounds like the drive spins up (one but not all? fully? partially?) and the system is frozen. Can't wake from sleep at all. If it does wake up and work, performance is too sluggish to use (molasses).

Cables and termination can be ruled out.

ATTO is brand new UL3S, updated to 1.6.4f1, and seems okay otherwise. One channel, two buses (meaning you could have IDs of zero on both internal and external, but all the IDs are unique and no ID conflicts).

I thought my frustration was over. Everything is/was fine UNTIL I tried going back to booting from internal drive on SoftRAID.

<finger pointers> Apple -> ATTO -> SoftRAID -> 9.2.1 -> B&W G3 (G4/500) -> Energy Saver... maybe the only "cure" is to forget Classic Mac OS 9.x and just go full time under 10.1 when it arrives next month.

Gregory

kaye
09-09-2001, 01:08 PM
Gregory,

I'm at a loss to see where your problem might be. I have used the UL3D with various configs of boot drives and RAIDs in both the PTP and QS (except of course I can't boot off of a striped RAID in QS). In the PTP, OS9.1, HFS or HFS+, ATTO firmware 1.50, 1.62, or 1.63. In the QS, OS9.2.1, HFS or HFS+, ATTO firmware 1.50 or 1.64f1. Both boxes, either G1 X15s or G2 X15s with SoftRAID 2.2.1 or 2.2.2.

Nothing sticks out about your config. k

[This message has been edited by kaye (edited 09 September 2001).]

TZ
09-09-2001, 03:31 PM
I wiped the drive and put Drive Setup 2.0.7, did low level format, running 9.2.1.

Quantum Atlas 10k II now has 9.2.1 on one volume and OS X on the other. But none of it is on RAID. Quantum + ATTO 1.6.4f1 on pre-G4 (but NewWorld) + SoftRAID + 9.2.1 just feels like there is something wrong. 30 yrs maintaining mainframes. Can't get a desktop up?

Maybe if I had done a low level with SoftRAID - but I had similar problems with Cheetah. And this is not on Cheetah, on drives that worked for months.

I'd have to go back to 9.1 and build a quick mirror and stripped raid. I'd like to hear from SoftRAID LLC that there actually is a change in 9.2 that needs to be addressed. For now, I definitely get good numbers from just Drive Setup (110MB/s sr/41MB/s sw).

At this point I am just getting the system stable and ready for 10.1 in the next month. If there is something more I should try first I will but I'd like to ask ATTO and just let them know what I am experiencing. What I wish is that Apple would be more up front so we don't have to wonder or guess, or vendors, if "maybe" an OS update impacts or affects some hardware or software or firmware we have.

The fact that sleep mode, Energy Saver, spining down disk drives, freeze on coming out of sleep and/or sluggish behavior point to bugs in the OS itself that need to be addressed. A house built on sand. I'm going to try and stick with OS X and avoid 9.2.1 at this point.


Gregory

TZ
09-10-2001, 02:38 PM
Beginning to look like some preference file or extension has been causing trouble. Won't rule out that even a cold boot can boot but fail. Low-level format just incase temperature changes had caused head drift (don't think that affects new drives but). And like someone else here, I'm using Atlas 10k II for booting. And the problem is more with internal than with any external. Foggy on if I ever had trouble with an external drive. There's no "log" of everything that has happened or changes in the last two weeks.

I'm good at telling people to do a clean install and not copy preference files and extensions over if in doubt. Or to move the Preference folder out of the system folder to quick-check if it might be causing trouble. ie, it happens only when menu begins to load before the clock is visible (first item always), or even Finder and volumes load.

I hear Apple is just getting information to 3rd party developers over changes in 9.2.1 that would impact SCSI devices. Ouch.

Gregory

TZ
09-10-2001, 06:53 PM
I'm back testing on 9.1

Seems that after shutting down for couple hours, even with the cleanest clean system running 9.2.1 on Drive Setup OR SoftRAID, the Finder doesn't come up, the menu bar didn't get as far as showing time, just froze... mouse moves around though.

So I'll know tomorrow if 9.1 is trouble.

All of these problems are booting from internal drives in B&W G3 (w/ XLR8 G4/500) off ATTO UL3S. A pair of Quantum Atlas 10k IIs. Both low levelled running Drive Setup on one and SoftRAID 2.2.2 on the other.

If I have trouble tomorrow, Mark James wants me to turn off Mode Page acceleration. After that... pull the ATTO and try the 2940U2B which never had any trouble with 9.1 or 9.2.1.

Want to remove SoftRAID? Option click while selecting "Quick Initialize" in SoftRAID.

I'm getting sick of this problem that looks like ATTO and my firmware.

Gregory

TZ
09-11-2001, 01:31 PM
9.2.1 is a NO. I have to use very minimal extensions and disable the Microsoft libraries it seems before reboot. 9.1 works. Tried every trick in the book I know.

Next step is to boot 9.2.1 from 2940U2B first with SoftRAID then with Drive Setup. Or to boot on external drives.

It is worse on cold boot but it does boot fine. But the drive or the problem... exists. Thinking there was "head drift" (not applicable nowdays is it?) I low level formatted my drives.

9.1 works. 9.2.1 is a problem.

Could 9.2.1 or ATTO have trouble with the XLR8 firmware that removes the G4 block on rev 2 B&W's?

gregory

TZ
09-11-2001, 08:13 PM
2940U2B works fine. Disconnected my drives from the ATTO UL3S (1.64f1) and attached to the 2940U2B. Boots 9.2.1 with full extensions etc w/o trouble or freeze or lock-up.

Not as good performance wise - but hey, no trouble booting OS 9.2.1 either. No conflicts. (knock on wood, I'll know for sure after it has been off, but I never had any trouble booting until I added the ATTO).

Bad combination? bad board? firmware? Apple supports the U2B in a way it doesnt the ATTO on B&W (w/ G4 block removed).

What do I ask ATTO when I email or call tomorrow?

Gregory

kaye
09-12-2001, 03:08 AM
Gregory,

Is there another machine that you can try the UL3S with the drive and cable/terminator in? Otherwise, I think there is a problem with the UL3S. Can you exchange the UL3S where it was purchased? Or get an RMA from ATTO. k

TZ
09-12-2001, 08:06 AM
kaye,

Zero problems booting on 2940U2B today. Never did that even once off the UL3S.

At first I thought - Blue G3 and mirrored RAID should work but hey, I'll split the mirror. Then removed SoftRAID and reformat low level with Drive Setup. Then that it was 9.2.1 and SoftRAID and/or the ATTO and its firmware.

I bought it from the MacGuru web store. Two weeks to rule out every combination. I don't have access to another machine, and if I did it would be a 7300 or G4/533, so different that I'm not sure what that would tell me.

Mark James has been talking to me and walking me through, hand-holding, priceless. Because I really thought Apple has changed the SCSI support in 9.2.1 and in Drive Setup 2.0.7. And Apple may have. Details on changes in 9.2.1 are only just now being released.

I contact Deb, but I think whatever happens ATTO needs to actually take the board into their labs.

Thirteen days of testing.

Gregory

TZ
09-12-2001, 06:05 PM
One interesting note: if I don't have any drives attached internally, I can boot from an external drive. It seems like the internal bus is faulty.

Gregory

kaye
09-12-2001, 09:32 PM
By itself, that last comment sounds like an ID # conflict between your internal drive/drives and your external drive/drives. I'm sure you have double checked that none of these drives have the same ID number. Time to get an RMA from Deb for an exchange. You could wait weeks while ATTO tests your existing board. k

TZ
09-13-2001, 10:01 AM
Kaye,

SCSI ID is definitely not involved. One channel, but drives on internal adapter show as one bus number, and external drives as another. All are on Granite cable and terminators and it only happens with 9.2.1.

Mark has said he can test etc so I may send him the card. Also, it may be something in the design of the UL3S and not show up in the UL3D. That would represent a $150 upgrade. Outside my budget for the next couple weeks.

Because it shows up under 9.2.1 I figured that it HAD to be a driver related issue. Or firmware. And maybe Drive Setup 2.1 or Mac OS 9.3 would solve this. With 10.1 shipping ~ October 1st I don't know if it would be important. But it does seem like an odd and yet major fault. But with who? Mac OS? SoftRAID? ATTO firmware? UL3S? and doesn't the UL3 use the same firmware for both 3S/3D?

I would hope ATTO has a database of all reports. Anyone with a UL3S, 9.2.1, etc with Blue G3. MAYBE G4 upgraded only?

Too many variables. And no reports of UL3Ds which got more testing and exposure.

Deb gave an okay to replace the card. I'd like to send it back, get a new UL3D, and then if she wants to send that card on to Mark for testing, while I'll send him a 10k II Atlas to use.

That's about the best I can come up with.

Gregory
(48 hrs of TV watching and counting)

kaye
09-13-2001, 01:07 PM
Gregory,

"SCSI ID is definitely not involved. One channel, but drives on internal adapter show as one bus number, and external drives as another. All are on Granite cable and terminators and it only happens with 9.2.1. "

Yes I realize that. I'm just suggesting that SCSI voodoo being what it is that you try unique ID numbers, given all the othe variables you have, such as 9.2.1, etc.

"Mark has said he can test etc so I may send him the card. Also, it may be something in the design of the UL3S and not show up in the UL3D. That would represent a $150 upgrade. Outside my budget for the next couple weeks. "

I'm not sure where Deb wants you to send the RMAed card but you will need to coordinate where you send it, Mark, or the Gurus Albuquerque address, or Hardware Haven, with a note inside about final destination for ATTO to test also. Getting complicated.

"Because it shows up under 9.2.1 I figured that it HAD to be a driver related issue. Or firmware. And maybe Drive Setup 2.1 or Mac OS 9.3 would solve this. With 10.1 shipping ~ October 1st I don't know if it would be important. But it does seem like an odd and yet major fault. But with who? Mac OS? SoftRAID? ATTO firmware? UL3S? and doesn't the UL3 use the same firmware for both 3S/3D?"

Beginning with ATTO firmware 1.6 and later, the firmware is the same for both UL3S and UL3D.

"Deb gave an okay to replace the card. I'd like to send it back, get a new UL3D, and then if she wants to send that card on to Mark for testing, while I'll send him a 10k II Atlas to use. "

That sounds like a good plan. Be sure to include a note in the box about where the card should go in what order so that each addressee knows where to send it next. k

TZ
09-13-2001, 01:34 PM
kaye,

Well, none of the drives ever had conflicting IDs.
Internal is 0, 1. External: 14, 12, 10.

The only thing I didn't do was to try using ATTO's driver.
That would be something they know, and should work okay
on their own board of course. They are going to do some
testing on their end with a UL3S and look more into 9.2.1.


So far it was only "preliminary" tests and reports regarding OS 9.2 compatibility. That from ATTO, FWB, and others.

Nested "IF" statements are I'm afraid "normal" anymore,)

Gregory

TZ
09-13-2001, 03:00 PM
Just formatted with ATTO ExpressPro Tools and it looks like issues with drives spinning up, booting from internal, are good so far. The real test is to shutdown overnight and full cold boot in the AM.

Gregory

magician
09-14-2001, 01:12 PM
sorry I'm late to the party.

so, it looks like it could actually be a SoftRAID issue after all, since it is behaving correctly when using ATTO Tools?

I myself have a B&W here that I use for testing, and it has a couple of X15's and a couple of 10k Cheetahs on it. This particular machine is booting off a RAID1 with just one problem, using a UL3D, SoftRAID 2.2.2, and OS9.1:

it freezes when it tries to load the desktop. I have to manually drop the machine into MacsBug, and ES quit the Formac TV control strip module (or something like that) and then resume to a full desktop. I went thru a million hours of troubleshooting, and got nowhere, but did find that nuking CarbonLib prefs is very prone to corruption.

the machine is otherwise very stable, and I run it for days between reboots. It wakes from sleep flawlessly.

So.

Not sure how pertinent any of this is.

Nelly is sitting in for Deb today, and received your email (she ran it past me, which is how I got vectored in here). We'll leave it for Deb to handle on Monday, since you still have a couple of days. We can upgrade you to a UL3D, exchange for another UL3S, or refund, at your discretion. The mechanics of all that will be up to her. That's why she's the Office Goddess.

hang in there.

http://macgurus.com/infopop/emoticons/icon_smile.gif

TZ
09-15-2001, 09:40 AM
magician,

I really appreciate the feedback and help. Clean installs, virgin system folder, 9.2.1 is only time I have trouble. I use CarbonLib 1.4.5.

I've learned that a first boot with shift key down "may" allow the 2nd boot to "clear" the desktop and make it to the Finder.

I was afraid that the freeze on menu bar was corrupting a preference or something and making it self-perpetuating.

ATTO tools may not resove the issue. ATTO says that there is hardly any difference in the UL3S/3D besides added channels (sounds like major difference to me). But has same firmware and controller perhaps.

I use to have some cold boots not "bong" and have to give it a second control - command - power button (ADB keyboard) booting off internal drives on the 2940U2B - so maybe the Atlas 10k II needs time etc.

My very first attempts etc were to use a new 2G X15 and 9.2.1 using Drive Setup 2.0.7. Not much help.

9.2.1 is "not intended for blah-blah" "is not a bug fix update" and "is not intended for all PowerMacs" yet it is, and starting with OS 6.0.5 with a long list of outstanding bugs from 6.0.3 and more bugs introduced in 6.0.4 I find Apple's OS updates, disclosure, a nightmare and frustrating. Last minute changes, breaks drivers w/o warning, undocumented APIs.

Yosemite issue? ATTO said they would look more closely, as did Mark (James). OS X 10.04+ doesn't seem to mind.

Gregory

TZ
09-17-2001, 05:02 PM
Just to update status on this problem:

<BLOCKQUOTE class="ip-ubbcode-quote"><font size="-1">quote:</font><HR>Dear Gregory,

An update on your situation. We have recreated this in our lab to little
success. Trying to boot off of a drive intialized with ATTO 2.6, or Apple
Drive Set-up 2.0.7 was unsuccessful. We hang after the OS loaded as well.

ATTO Technology
Technical Services<HR></BLOCKQUOTE>

I am both relieved - and partially shocked.

Gregory