Kernel Panic In Mac OS 10.6.5

See updated post, it wasn't 10.6.5.

Anyway I did try to install the 10.6.4 combo but it failed several times, then it KP.

Must be memory, or board? Can it be the dvd drive, even if nothing is in the DVD? The reason I say that is that the re-installation got stuck a few times.

Every Mac ships with installation/restoration discs. Those also include the AHT, even if it's preinstalled on the machine. Call AppleCare and request replacement discs. Alternatively, visit an Apple repair center.

The MacBook shipped with OS X and an Apple Protection Plan CD. Tomorrow I'll check that CD (it's in my office).

However I'm running Rember, and it's finding errors:

FAILURE! Data mismatch at local address 0x000000013c593488
Expected Data: 0x0af3e8b4621b183c, Actual Data: 0x0afbe8b4621b183c

Should I just go ahead and replace the memory or go to a repair center? The MPB is out of warranty. I think that repair center will charge $200 to fix any problems.

Thanks for the replies by the way.

Look at the OS X disc, there should be a mention of AHT on it. It's not on the AppleCare protection disc.

If Rember's finding errors, then there's issues with the installed RAM. If Apple's, they should replace it, call AppleCare. If 3rd-party, then it should be lifetime. Replacing RAM is a user option.

Check the MBP manuals that pop up with a google search for *mbp manual site:apple.com*.

Ok Baitwo you definitely rock, thanks for all your help.

I think that what happened is that I got rid of the original DVD since I bought Snow Leopard retail. Ok, it might seem very stupid, but it was my first experience as OS X user and didn't know about AHT (I came from Linux/Windows)

I'm playing with Rember and memory chips / slots. One of the two is bad, Rember takes some time to run, I'll soon know the answer. If it's the chip, I guess replace the chip, otherwise it's the controller or motherboard I guess.

The memory is 2 x Samsung 2GB 2Rx8 PC3 8500S .

What do you mean lifetime? I can send it to Samsung to get a replacement?

Most likely the RAM module. If it were the controller or motherboard, none would work.

Correct, it was one of the RAM modules.
Removing seems to have solved the Kernel Panic, running since 7+ hours now w/o troubles.

Will buy another 2GB module tomorrow.

I went back to 10.6.4 and everything is good again.

10.6.5 seems to have some serious issues - I've found many descriptions online that are just like my experience: When the OS goes to sleep, it doesn't wake up. Instead you get a black screen and a hard crash, only a hard reboot can recover it.

Here's how I went back to 10.6.4 with minimal trouble:
- Found my Snow Leopard install disk
- Installed Snow Leopard OVER the bad 10.6.5 install
- Applied 10.6.4 update

Now everything is back to normal and all user preferences, data, etc have been preserved. Not going back to 10.6.5 until it's fixed.

I have upgraded my system with 3rd party RAM (8GB) but I don't see how that could cause the problem given that this RAM works flawlessly in 10.6.4 and before. Would 10.6.5 place more stringent requirements on RAM? Very unlikely! My best guess is that it has something to do with those graphics improvements and the integrated NVidia chips on the 2009 MBPs. I could have tested this by turning on the dedicated graphics card, but ... life's too short. I already wasted inordinate amounts of time on this.

AFAIK, each subsequent update pushes the OS into meeting its stringent requirements. If you don't exhaustively test the RAM, you just making WAGs.

@baltwo - may I ask where you get this information from? I find it hard to believe.

RAM timing, and RAM access are controlled at a Firmware (EFI) level. This level is below the operating system, closer to the hardware. The operating system doesn't have anything to do with that - it's just using services that the EFI provide it. I find it therefore highly unlikely that any OS update could have the effect of making RAM requirements more stringent - only a Firmware (EFI) update could do that.

EFI updates are issued occasionally by Apple, but relatively rare. OS X 10.6.5 does not update the firmware. Firmware updates also can't be rolled back.

RAM problems are tricky to determine because they will often show random behavior - but I think the fact that all issues went away when I downgraded to 10.6.4 effectively rules out any hardware problem. The issue is isolated to the 10.6.5 update, and the software side.

Message was edited by: orthorim

orthorim wrote:
RAM problems are tricky to determine because they will often show random behavior - but I think the fact that all issues went away when I downgraded to 10.6.4 effectively rules out any hardware problem. The issue is isolated to the 10.6.5 update, and the software side.

You absolutely cannot make that assessment.

If 10.6.5 has a slightly different access pattern to RAM it can run across issues 10.6.4 does not (and vice-versa.)

I've seen this time and again when debugging systems where RAM shows as bad only under certain circumstances and in certain access patterns; running a comprehensive RAM test is one of the only ways to rule RAM in or out, and even then not until you've swapped around DIMMs and have rerun the test.

I speak from my experience.

I really thought it was 10.6.5 update causing the kernel panics. But it wasn't.
It was a bad memory module.

My guess is that 10.6.5 used memory differently. Maybe it put more strain on the ram. No idea. The net result is that the kernel panics happened 99% on 10.6.5. The ONE time that happened on 10.6 made me realize it wasn't the OS version.

When I identified the hardware problem and solved it, I could instal 10.6.5 and I'm running w/o troubles now.

My somewhat educated guess is that you're experiencing the same symptoms. Try Rember to test your memory modules.

I have a MacBook Pro 5,3 (as listed in ASP) I have had 3 kernel panics and 2 blue screens on startup since installing the 10.6.5 update. I have run Rember for the RAM testing (all ok) zapped the PRAM and NVRAM and repaired permissions from 10.6 install disk (a number of permissions issues mainly in JavaVM framework). I've checked on campus and there are multiple users experiencing the same issue.

Of the kexts loaded all bar Sophos 7.2.0 are apple ones, the Airport BRCM43xx is the only kext unloaded so nothing untoward there.

What has been said about RAM is quite right in that some system updates may highlight problems that didn't appear to manifest before but it is not always definitively the case in this instance.

I then disconnected the USB cable that runs from my Apple 20" which itself has a LaCie HD and my Apple keyboard and mouse attached. Having done all this no panic. Re-connected my Apple monitor minus the HD, no panic. Connected the same HD via FW800 to the MBP still no panic. Now this isn't an exact definite cause in that there are others experiencing the same problem but it does appear that those experiencing it do use at least some form of USB peripheral.

FWIW, the JavaVM permissions items are known and should be ignored, since DU can't repair symbolic links when expecting files.