Description:
OSX server crashdump is always identical:
Exception Type: EXC_BAD_ACCESS (SIGBUS)
Exception Codes: KERN_PROTECTION_FAILURE at 0x0000000000000104
Crashed Thread: 3
Thread 3 Crashed:
0 libSystem.B.dylib 0x93c2b38d _dispatch_wakeup + 91
1 libSystem.B.dylib 0x93c2ae96 _dispatch_source_create2 + 194
2 libSystem.B.dylib 0x93c2d72b dispatch_source_machport_create + 60
3 ...ple.CoreServices.CarbonCore 0x94fdb21f _initSharedCache() + 131
4 libSystem.B.dylib 0x93c25090 pthread_once + 82
5 ...ple.CoreServices.CarbonCore 0x94fbe83c scCreateSystemServiceVersion + 82
6 ...ple.CoreServices.CarbonCore 0x94fdb005 FileIDTreeGetCachedPort + 299
7 ...ple.CoreServices.CarbonCore 0x94fbf89c FSNodeStorageGetAndLockCurrentUniverse + 39
8 ...ple.CoreServices.CarbonCore 0x94fcddb9 FileIDTreeGetVRefNumForDevice + 29
9 ...ple.CoreServices.CarbonCore 0x94fcdd32 FSMount::FSMount(unsigned int, FSMountNumberType, short*) + 62
10 ...ple.CoreServices.CarbonCore 0x94fcc1b4 PathGetObjectInfo(char const*, unsigned long, unsigned long, short*, unsigned long*, unsigned long*, char*, unsigned long*, unsigned char*) + 314
11 ...ple.CoreServices.CarbonCore 0x94fcbfb4 FSPathMakeRefInternal(unsigned char const*, unsigned long, unsigned long, FSRef*, unsigned char*) + 134
12 ...ple.CoreServices.CarbonCore 0x94fcbf2c FSPathMakeRef + 47
13 com.apple.CoreFoundation 0x91cd7e74 __CFCarbonCore_FSPathMakeRef + 68
14 com.apple.CoreFoundation 0x91cd7bbf _CFGetFSRefFromURL + 767
15 com.apple.CoreFoundation 0x91cd78b2 CFURLGetFSRef + 34
16 com.apple.CoreFoundation 0x91cd76eb _CFBundleCopyInfoDictionaryInResourceForkWithAllocator + 75
17 com.apple.CoreFoundation 0x91cd7694 _CFBundleCopyInfoDictionaryInResourceFork + 36
18 ...ple.CoreServices.CarbonCore 0x94fd18df GetBugsForOurBundleIDFromCoreservicesd + 568
19 ...ple.CoreServices.CarbonCore 0x94fd164c _CSCheckFix + 20
20 ...ple.CoreServices.CarbonCore 0x94ff6e72 _Gestalt_SystemVersion + 781
21 ...ple.CoreServices.CarbonCore 0x94ff6753 Gestalt + 162
22 libsigar.0.dylib 0x0233e4ba sigar_os_sys_info_get + 154
23 libsigar.0.dylib 0x0232eb29 sigar_sys_info_get + 57
24 libagent.dylib 0x0034f30f agent_dc_os_os_update_values + 63
25 libagent.dylib 0x0034c3b6 job_collect_get_value + 214
26 libagent.dylib 0x003506c6 job_collect_os_thread + 246
27 libglib-2.0.0.1600.6.dylib 0x000df8de g_thread_create_proxy + 174
28 libSystem.B.dylib 0x93c33fbd _pthread_start + 345
29 libSystem.B.dylib 0x93c33e42 thread_start + 34
oddly agent-run-os-tests is fine calling the same sigar_sys_info_get:
sigar-test-all.c.554 (test_sigar_sys_info_get):
sysinfo.names = MacOSX
sysinfo.versions = 10.6.2
sysinfo.archs = i386
sysinfo.machines = i386
sysinfo.descriptions = Mac OS X Snow Leopard
sysinfo.patch_levels = unknown
sysinfo.vendors = Apple
sysinfo.vendor_versions = 10.6
sysinfo.vendor_names = Mac OS X
sysinfo.vendor_code_name = Snow Leopard
This would be our 2.1.1.1144 agent build which is 32bit running on osx server 10.6.2. Agent crashes, then receiveds a dupuuid from the service manager, running in log-level=debug shows:
2010-02-26 12:56:43: (debug) job (task 33016 (collect_os)) executes again in 300 sec
2010-02-26 12:56:44: (debug) chassis.c:282: 3139 returned: 3139
2010-02-26 12:56:44: (message) chassis.c:304: [angel] PID=3139 died on signal=10 (it used 0 kBytes max) ... waiting 3min before restart
2010-02-26 12:56:46: (message) chassis.c:259: [angel] we try to keep PID=3146 alive
---
2010-02-26 12:59:13: (debug) scheduler.c.520: scheduling collect_os for os::net->netmask
2010-02-26 12:59:14: (debug) chassis.c:282: 3146 returned: 3146
2010-02-26 12:59:14: (message) chassis.c:304: [angel] PID=3146 died on signal=10 (it used 0 kBytes max) ... waiting 3min before restart
2010-02-26 12:59:16: (message) chassis.c:259: [angel] we try to keep PID=3183 alive
How to repeat:
Unable to repeat as we don't have access to a '10.6.2 osx server', am unable to repeat on normal osx 10.6.2 client.
Suggested fix:
Needs additional testing specifically on osx 10.6.2 server version. Our builds are 10.5 32bit and assumed to work ok.
