Who likes major diagnostic challenges?

[Benbikeman]

I'm tearing my hair out here, as I have a problem which has some symptoms which suggest it must be software and some symptoms which suggest it must be hardware. Despite two visits to Apple Stores and numerous calls to AppleCare, we're no closer to a diagnosis a week later.

Setup:
MacBook Pro 17 with 2 x 1Tb drives (drive 2 in DVD bay)
Booting from drive 1 in main bay
OSX 10.8.1

Original symptoms:
Kernal panics, followed by boot volume on drive 1 being intermittently mountable

What we have done so far:

Replaced physical drive
Clean install of OS X, and did a migration of everything except settings
Same symptoms

Clean install of OS X on drive 2
Runs perfectly with nothing else installed

Put original drive back in drive 1
Clean install of OS X on that drive
Runs perfectly with nothing else installed

Created a partition on drive 2, clean install of OS X in that partition
Migrated everything back to that partition
Same symptoms when booting from that partition
NB: Different physical drive, different physical drive bay
This suggests software rather than hardware, but it doesn't end there ...

On drive 1:
Reinstalled Thunderbird (can explain why this and not Mail if anyone cares)
All ok

Manually copying data ONLY from Time Machine then gets crashes:
Thunderbird first
Then Safari
Neither window disappears even when force-quit processes
Crash Report showing 100% CPU but no disk activity
Any app opened in that time would run but not quit (eg. Activity Monitor)
Shutdown did nothing, had to do a hard shutdown

Booting into Utilities, Disk Utilities can't unmount the volume on drive 1

Review so far:
Clean install on either drive runs fine with both Thunderbird and Safari installed
But both apps crash when data is copied back
Drive 1 mount problems return when the data is back on it
BUT the data is just that, documents only, no apps or settings

What I want to do at this stage is erase drive 1 again, clean install OSX again, run Mail rather than Thunderbird, make sure that system is stable, then copy data back (manually) then see if that is stable, but it's all feeling rather trial-and-error.

Anyone have any bright ideas?

Ben

 

[ljonesj]

sounds like the info for thunderbird data is corrupted that is causing panics

----------

sounds like the info for thunderbird data is corrupted that is causing panics

 

[Benbikeman]

sounds like the info for thunderbird data is corrupted that is causing panics

Which fits many of the symptoms, yet ... it runs fine without my documents restored. I'm literally pulling back documents only, no system files, no app files. This is what's baffling me.

Ben

 

[Benbikeman]

Finally solved!

It was faulty RAM.

Because the initial symptom was the boot volume refusing to mount, it was initially pointing to the drive or drive-related hardware. When the same symptoms occurred when migrating to a new drive in a new bay, we then thought software.

But I think what happened is that the first crash corrupted something on the boot volume in such a way that the corrupted file was also migrated back. The really odd thing was a vanilla install running fine for a while, but that now appears to be coincidence.

Memtest reported shedloads of errors on the RAM. I swapped back to the original 4Gb and ran it again, getting a clean result. The machine has now been running for 12 hours with the original RAM without a crash (crashes were previously happening once every 1-2 hours).

Now I have a clean install on drive 2, I shall migrate back to drive 1 and just see whether that works. It may be there was some weird combination of a corrupted file and bad RAM causing the crash on the new drive, so we shall see. Worst-case, I reinstall apps and settings manually on drive 2, clone that to drive 1 and am back in action.