This weekend I’ve realised that I’m finally in a state of Mac hell. Here’s what tipped me over the edge.
As I’ve described in the podcast, I’ve set up a striped RAID array with two firewire disks connected to my Mac Mini and set that as the Time Machine disk. This being a machine that’s almost permanently on, I’ve set up my other Macs to use this as a Time Machine drive over AFP. While it was supremely slow over 802.11g wifi, it worked.
Since the Leopard upgrade my Mac Mini has kernel panic’d about 6 times, all whilst being basically idle (except perhaps for Time Machine). To isolate the problem I took away the RAID array, reformatted one drive and set that as the Time Machine drive temporarily. Now Time Machine crashes every time it attempts the initial backup, every time at about 120MB in. It doesn’t present an error to the user. It just silently dies. Here’s a snippet of the crash reporter log:
Process: backupd [226]
Path: /System/Library/CoreServices/backupd
Identifier: backupd
Version: ??? (???)
Code Type: X86 (Native)
Parent Process: launchd [1]
Date/Time: 2007-11-11 09:29:54.987 +1100
OS Version: Mac OS X 10.5 (9A581)
Report Version: 6
Exception Type: EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x0000000002678d30
Crashed Thread: 3
……
Thread 3 Crashed:
0 com.apple.LaunchServices 0x908a5fe4
LSHandlerPref::roleHandler(unsigned long) const + 36
1 com.apple.LaunchServices 0x90894296
_LSEvaluateClaimArray(LSBindingState*, unsigned long const*, unsigned
long) + 268
2 com.apple.LaunchServices 0x90893dc3
_LSGetBinding(LSBindingState*) + 249
3 com.apple.LaunchServices 0x90893c80
_LSGetBindingStateForNode(LSBindingState*, FSNode*) + 484
4 com.apple.LaunchServices 0x90893057 _LSGetBindingForNode + 149
5 com.apple.LaunchServices 0x9087c667 _LSCopyInfoForNode + 2457
6 com.apple.LaunchServices 0x9089849a _LSCopyItemInfoForRefInfo + 124
7 com.apple.DesktopServices 0x90ec4a66 TFSInfo::FetchData() + 82
8 com.apple.DesktopServices 0x90ec5540 TFSInfo::IsApplication()
const + 18
9 com.apple.DesktopServices 0x90ec2bc0 TCopyWriter::GetNextItem() + 344
10 com.apple.DesktopServices 0x90ec28f4 TCopyWriter::Write() + 102
11 com.apple.DesktopServices 0x90ec2482
TCopyWriter::WriteTaskProc(void*) + 70
12 ...ple.CoreServices.CarbonCore 0x965754bb PrivateMPEntryPoint + 56
13 libSystem.B.dylib 0x9424c075 _pthread_start + 321
14 libSystem.B.dylib 0x9424bf32 thread_start + 34
There are numerous threads in the Time Machine support forum with similar problems where the solution reminds me of the time honoured Windows solution… reboot. None of these worked for me, and the fact is that this works on my MacBook at work with a different drive over USB. A firewire issue perhaps?
The Mac Mini no longer kernel panics by the way.
Now what other problems have I had to have this tip me over the edge? Let’s see:
- My MacBook sometimes doesn’t go to sleep
- When it does go to sleep it doesn’t wake from sleep half the time
- Bluetooth drops out all the time and when it works it seems a bit dodgy (people complain about sound being scratchy when I use my bluetooth headset with Skype; works fine with my mobile phone)
- I had the infamous “Battery not available” bug that mysteriously went away (the official update didn’t fix the problem initially)
- My MacBook has a dodgy display, one that my wife’s Core 2 Duo or Peter’s 1.6Ghz MacBook doesn’t have
The list continues and it makes my heart break as I’m a big fan of Apple and Mac OS X.
Granted I am an abnormal user, with my mac being my development environment with lots of my own background processes running most of the time, but I never had such problems on PowerPC hardware. I think that Apple are still struggling with the move to x86 given that you see firmware updates being released almost every month.
Bring back the good old days I say!
Update: 10.5.1 does not fix my time machine problem.
