Kernel panics, disc errors, general bad times

[Achilles75]

Lengthy post alert.

I switched over from a PC to a MBP last fall because I was tired of the instability of PCs and Windows. Bought a 15” early 2011 MBP, 2GHz, 4GB RAM, 500GB HDD, Lion 10.7.2. Promptly upgraded the RAM to 8GB. Machine ran fine for a couple of months; very pleased.

Here are the highlights of my troubles:

A couple of weeks ago I started getting the occasional gray screen, need to restart kernel panics. They increased in frequency until a couple of days ago when I couldn’t reboot. When into the recovery partition and found a “keys out of order error – rebuilding catalog B-tree, volume Mac HD cannot be repaired. Crap.

Called Apple. Fortunately (or not – we’ll get to that part), I have kept Time Machine backups on an external HD. So at Apple’s suggestion I erased Mac HD and restored my system from Time Machine. I backed up a few weeks to exclude a few new programs that I was a little suspicious of. Anyway – restore went fine, machine started booting again normally.

Overnight, I attempted to create a brand new Time Machine backup on my external HD. I awoke to a kernel panic/gray restart screen. I restarted. I tried to verify discs – my external TM disc gave an invalid node structure/cannot be repaired. My Mac HD gave an invalid key length. Was able to repair the Mac HD with disc utility back to “volume appears ok” status.

At this point, I became concerned that my external hard drive I’d been using for Time Machine was perhaps bad. So I bought another external HD. Formatted property and started creating yet another TM backup. Had two more kernel panics while trying to create the backup on the new external.

Called Apple. Reset PRAM. Was told that might eliminate the panics. Continued trying to create a new backup image. Another panic. Called Apple. Suggestion was to reinstall Lion. Not ready to reinstall OS yet. This is why I bought an Apple in the first place – stability. Crap.

This whole time, I have also been suspicious of my Chrome browser not playing nicely, so I did kill Chrome and remove all of its pieces and parts from my system. After removing Chrome, I nuked my new (2nd) external HD and started again to make a new backup in Time Machine. Well, it worked. Created the whole backup with no kernel panics. System still up and running.

However – this morning, again on a whim, I ran a verify disc on the new TM backup, and guess what? Invalid node structure, needs to be repaired. And on my Mac HD – invalid key length, needs to be repaired. Of course, now, I know what will happen when I try to repair it – disc utility won’t be able to fix it the external drive and it will become unusable again. And I will be able to repair Mac HD and it will be fine for an hour or two, but the next time I try to verify it, there will be another error.

So basically, I have an unstable system which I can’t seem to backup with Time Machine. I’m currently copying over my actual data to yet another external disc just to save it. I suppose the next step is to reinstall Lion. But I wanted to get some opinion here first. I bought a Mac to get away from this sort of crap, and I’ve now wasted the better part of 3 days on a very PC-like experience of instability.

 

[KnightWRX]

Seems more like a hardware problem. Bad RAM is probably the most likely culprit.

 

[Achilles75]

Seems more like a hardware problem. Bad RAM is probably the most likely culprit.

I honestly don't mean this to sound snarky, because I'm in over my head here and need some help/suggestions. I probably know a little more than most computer users, but I'm not a techie by a long shot. And this is my first Apple machine.

Why haven't any of the multitude of folks I've talked with at Apple Care over the last few days suggested a hardware problem? Why would the RAM work fine at first and gradually crap the bed? Why would faulty RAM be trashing my external HDs? What about my scenario makes you immediately think bad RAM rather than something else?

And my big question - when the console constantly generates system diagnostic reports and kernel logs, why isn't Apple asking for them? Can't these diagnostics pinpoint a problem, or at least point us to software vs. hardware issues?

 

[KnightWRX]

Why haven't any of the multitude of folks I've talked with at Apple Care over the last few days suggested a hardware problem? Why would the RAM work fine at first and gradually crap the bed? Why would faulty RAM be trashing my external HDs? What about my scenario makes you immediately think bad RAM rather than something else?

Bad can result in any number of corruptions. Remember, your Kernel and userland applications all run in RAM. If data in there gets corrupted, bad things get written to disk caches which then get flushed to disks and if code portions of the kernel get corrupted or become unaccessible due to a faulted page, well, panics occur.

Not to mention repeated Kernel panics caused by said bad ram can result in unflushed writes which can mean filesystem corruptions if some of these writes were partially committed.

Why the myrriad of Apple employees haven't suggested it ? I dunno. But you said you upgraded the RAM, did you buy Apple branded RAM installed by Apple or did you buy from a 3rd party ? Do you still have the original RAM ?

I'd try that. But that's me.

 

[Achilles75]

But you said you upgraded the RAM, did you buy Apple branded RAM installed by Apple or did you buy from a 3rd party ? Do you still have the original RAM ?

I'd try that. But that's me.

3rd party RAM, and yes, still have original Apple RAM.

System has seemed stable (no kernel panics) since killing Chrome. (I know 16 hours isn't much time, but considering I had 5 kernel panics in the preceding 24 hours, I'm taking this as a good sign.)

Right now I've got my fingers crossed on that front before I swap the RAM out again or go re-installing the OS.

 

[tyche]

My story is on Windows but I had upgraded my PC from 4gb to 8gb and everything looked fine for some time (windows 7 64bit). However, I would get file corruption all the time. In particular, .zip, or .iso would work one day and then the next it was saying the file is corrupted. This had be baffled for a long time as I did everything I could think of to fix it. Finally, it dawned on me I added extra memory and pulled it out. Presto, no more file system corruptions.

You could be getting this, when your system accesses the high memory which might not happen until you load the system.

I would run the machine with just the original Apple memory for as many days as you think necessary to get it to fail on you to rule this out.

 

[Achilles75]

I appreciate the thoughts. Just called Apple again - suggestion now was to restart in safe mode (this somehow rewrites boot directories?). Afterwards, nuke and pave both external hard drives I've been using. Drag and drop data to one, Time Machine backup to the other. If they both manage to write without becoming corrupt, I'm ok. If they become corrupt again, nuke and pave Mac HD and do a clean reinstall of Lion. So I suppose if *that* doesn't do it, the RAM would be the culprit. I imagine I will try swapping the RAM before a clean install of Lion. I specifically asked the last guy I just spoke with at Apple about the RAM and he didn't seem to suspect it. He's the second one who said sounds like software problem somewhere along the line.

In the meantime, I'm creating backups (again) to see if they can be written without disc errors.

 

[Achilles75]

Seems more like a hardware problem. Bad RAM is probably the most likely culprit.

You know what? I think you were right. Yanked the 3rd party RAM and things seem to be running ok.

 

[blackburn]

I bet on ram issues too. I've had corrupted hard drives and hangs on my old Pentium 3 and it was a bad ram stick. You can run apple hardware test to check if the ram is ok.
People should *always* test their machine with some kind of stress test after upgrading something. It will save a lot of trouble down the road.

 

[Gomff]

Another vote for RAM here. Kernel Panics, especially that many are usually indicative of bad hardware. First time I experienced them my GPU was on it's way out, the second time it was RAM.

Hope you get it sorted out.

 

[Achilles75]

I bet on ram issues too. I've had corrupted hard drives and hangs on my old Pentium 3 and it was a bad ram stick. You can run apple hardware test to check if the ram is ok.
People should *always* test their machine with some kind of stress test after upgrading something. It will save a lot of trouble down the road.

We tried to erase the Mac HD and do a clean install of Lion and it wouldn't do a thing. Apple was sending me a box for repair; the final tech I talked with was suspicious of a failed hard drive or logic board. And still didn't suspect the RAM. Before mailing the machine off for repair, I swapped the RAM out for the original just out of curiosity, ran a clean install, and suddenly everything is fine again.

For future reference, as a new Mac owner, what is this "stress test" and how do you test fire Apple hardware components?

And two, where do you guys buy your aftermarket RAM? I know anyone can have a faulty stick, but the more I read about it now (after the fact, of course) it seems that Crucial has some issues in newer MBPs.

 

[KnightWRX]

And two, where do you guys buy your aftermarket RAM? I know anyone can have a faulty stick, but the more I read about it now (after the fact, of course) it seems that Crucial has some issues in newer MBPs.

I had no luck on my 2008 Unibody MacBook with Crucial either. My GF's white MacBook is running fine with it, but 2 batches were causing mine to not wake up from sleep half the time. I finally put mushkin RAM in it and everything was fine. My home server is running PNY (4x4 GB) and it's been flawless.

As for testing RAM, really, there isn't much you can do. Even running memtest for hours on end, everything can be peachy with it and after 3 months, the issues start. Or right away. Or whenever. RAM is finicky and unfortunately, unless you have ECC memory, when those bits flip, bad things will happen.

 

[Gomff]

I've had mixed results with Crucial in the UK at least.

Laptop memory on a fairly old MBP was fine, but ECC memory on a Mac Pro had to be returned twice before I finally received error free memory. In fairness to them though, they were very apologetic and replaced it with no quibbles.

I think the guarantees should be at least a year, so you have plenty of time for it to bed in. Buying memory from anywhere can be a bit of a lottery.

As far as memory testing goes, I've used Rember before and found it adequate and free.

 

[blackburn]

I've used corsair ram with 0 problems. I usually do either a day of memtest or apple hardware test.

 

[thekev]

Backups should be started before you have problems. If you need to test ram, memtest is popular. Run five cycles in single user mode. Bad ram can cause errors to accumulate over time. Where did you buy your ram? I always buy something that's at least reported to work on that specific model even though the Intel Macs are much less picky than the PowerPC generation that they replaced.

 

[Achilles75]

Backups should be started before you have problems. If you need to test ram, memtest is popular. Run five cycles in single user mode. Bad ram can cause errors to accumulate over time. Where did you buy your ram? I always buy something that's at least reported to work on that specific model even though the Intel Macs are much less picky than the PowerPC generation that they replaced.

Crucial from Amazon.

 

[thekev]

Crucial from Amazon.

I'd just run memtest in single user mode to check it. Any ram can turn out bad. Let it run 5 cycles or so overnight. If either stick generates any errors, return it. You can also test with the original ram. The reason it has been suggested is that you're exhibiting classic bad ram symptoms.