Q: Dual Quad-core Mac Pro reboots from sleep

Okay, that's fair enough. Apple Tech Support merely lied to me. Hope of hopes. Though why they should both additionally (and independently) speak (lie) about an imminent fix for BTO machines is mysterious.

Let's hope / assume they are doing their best with what they have been told so far - and that a fix (if software...) is forthcoming and will in fact be for stock and CTO systems.......

probably it's not cool to say "we have no idea if they are trying to fix it or not".

I think I may have a fix for this but so have a lot of other people, including me, only to have the problem reappear. My latest 'fix' is to run Leopard Cache Cleaner's Deep Clean option. I will give it a few more days before I am confident that this will do the trick.

Well, so much for my fix (Leopard Cache Cleaner's Deep Clean.) After a number of successful wake-from-sleeps varying in time between 15 minutes to several hours I decided to test the system by putting the system to sleep and as soon as the power light faded a few times indicating sleep mode I woke the system up with the space bar at which point the fans started to roar (on the Mac Pro it's not really that much of a roar) and the system hung.

I wish Apple would make some sort of official statement about this problem, i.e. under investigation, known problem etc.

jwholder: How long did the fans roar? Did you give it sufficient time? I'm asking because I had a similar experience, and was just about ready to give up when the screen appeared. It wasn't a restart, but it seemed to take longer than usual to wake. I've only had the restart on wake one time, but I did have the fans go crazy another time after an Adobe software update and restart.

MacfixIt article on the subject, linked below, summarizes the issue nicely:

http://www.macfixit.com/article.php?story=2008022111294810

Chateaubugs wrote:
MacfixIt article on the subject, linked below, summarizes the issue nicely:

http://www.macfixit.com/article.php?story=2008022111294810

But, as has been the case here, that article also contains misinformation (and in any event, I think it's mainly drawn from messages already posted on this very thread).

People have made suggestions as to cause that just don't make sense, e.g.,
1. inadequate power provided by a UPS (most of us don't have our machines plugged into a UPS)
2. running "deep cache cleaner"
3. it's CAUSED by the graphics update (the problem was reported on vanilla machines long before the graphics update was posted)
4. "it's software/no it's hardware/no, it's software" (reminds me of the long ago fake TV commerical from Saturday Night Live: "It's a desert topping/no, it's a floor wax"). Apple must know SOMETHING by now, yet their willingness to offer some customers new machines that apparently also suffer the same problem suggests they may not, or that if they do they're being disingenuous in their response.

In the end, as I've said earlier, the (thus far) uncontested report that the problem does NOT occur while Mac Pros are running Vista only natively (via Boot Camp, not via virtualization) may be the best evidence that it's a software problem (or at least that it's rooted somehow in the manner in which the Mac OS and Windows address power management in the hardware).

Most people, anyone with a 865W Mac Pro, should have a UPS, many do, and may need to have it on its own circuit. The old Mac Pro had one big problem, that may still be around, the inrush current on wake from sleep may only register 400W or so, but would trip a system and cause problems.

That is one of the items that Mike ( www.xlr8yourmac.com ) asks when people submit a report to check for.

I think newer models have a later revision of the motherboard and firmware.

I have always used Cocktail to resolve any issues on my G5 and my new MP. As soon as my MP restarted from sleep, I installed Cocktail and started using it and haven't experienced the problem since.

The hatter wrote:
Most people, anyone with a 865W Mac Pro, should have a UPS, many do, and may need to have it on its own circuit. The old Mac Pro had one big problem, that may still be around, the inrush current on wake from sleep may only register 400W or so, but would trip a system and cause problems.

If you go to an Apple Store, you'll see them plugged into ordinary a/c circuits (not even a surge protector. I'd wager fewer than 20% of those in the chorus of complaints on this thread have a UPS. In any event, the notion that an inadequate UPS is the cause is dispelled by even one report that it happens on machines not plugged in through a UPS.

I think newer models have a later revision of the motherboard and firmware.

I've heard a few people claim this, but haven't seen any documentation of that whatsoever, and there are many people opening new boxes within the past week or two, plugging them in, and encountering the reboot from sleep the very next morning.

As far as a firmware revision is concerned, anyone have any idea how we could find out if this is the case? If it were true, I'd think Apple would announce they had a fix rather than permitting this ever burgeoning thread continue to embarrass them.

Shaken or stirred??

Or how about a DNA centrifuge? 'cause the DNA really needs to be separated.

I don't know when a 'shipping' system was 'built' which they could determine. And then run Windows or find the motherboard DNA.

Only a handful that were asked to send it in probably. And it wouldn't be the first time there are different systems, some liquid cooled G5s had totally different parts. Crucial tracks memory closely and they revise their production and engineering along the way, sometimes with good results, but sometimes not.

If you go to an Apple Store, you'll see them plugged into ordinary a/c circuits (not even a surge protector. I'd wager fewer than 20% of those in the chorus of complaints on this thread have a UPS. In any event, the notion that an inadequate UPS is the cause is dispelled by even one report that it happens on machines not plugged in through a UPS.

Not exactly. A UPS can cause it own issues that may look superficially the same. Only by removing the UPS can you discern its influence.

I've heard a few people claim this, but haven't seen any documentation of that whatsoever, and there are many people opening new boxes within the past week or two, plugging them in, and encountering the reboot from sleep the very next morning.

A careful reading of the threads does not show any trends of early vs late. There seems to be an underlying failure rate that has been constant...

islk350 wrote:

If you go to an Apple Store, you'll see them plugged into ordinary a/c circuits (not even a surge protector. I'd wager fewer than 20% of those in the chorus of complaints on this thread have a UPS. In any event, the notion that an inadequate UPS is the cause is dispelled by even one report that it happens on machines not plugged in through a UPS.

Not exactly. A UPS can cause it own issues that may look superficially the same. Only by removing the UPS can you discern its influence.

Perhaps I wasn't emphatic enough in my previous message. I said a UPS couldn't be the cause; I'm optimistic that this is a gremlin to which Occam's razor can be applied, and that there is indeed a single cause of the problem. Hence, the fact that the wake from sleep reboot happens on machines not attached to a UPS rules out the UPS as the cause.

If there is a hardware issue, perhaps a UPS could make it more likely to appear, but my reading of the posts from people who have them seems to indicate the gremlin tends to surface after long sleep intervals, not happen after short sleep cycles, which is what most people who don't have UPSes are experiencing as well, so I've not seen anything that suggests different behavior experienced by people who have UPSes vs. those who don't.

And I'm not ready to characterize the gremlin in terms of a "failure rate." I know some people report they've never seen the problem, but my money is on some bug in the power management part of the operating system that gets triggered by sleep cycles. Perhaps some of the voodoo and incantations that people have reported as stopping the gremlin (usually temporarily, unfortunately) interact somehow with a bug in the OS. I'm not sending my machine in because it doesn't misbehave as long as I don't sleep the entire system. I don't really know how much power it consumes when the drives have spun down and the displays are sleeping, but it's certainly not generating much heat then, and my data has been safe, which is what I REALLY care most about.