TALIGENT TOOLS FOR AIX - System crawls and CommonPoint eventually hangs

System crawls and CommonPoint eventually hangs

While starting the layer, SOL begins to load the libraries and then the servers start up. This process is a crawl compared to normal speed, and then one or more of the servers receives a SIGILL and it is downhill from there. You might see a listing like this when you start the system:

...
# Loading libraries...
# Starting merged servers in 3 clumps ...
# Starting service thread: NameServer.
# Starting service thread: TokenServer.
...
Dumping core after receiving signal SIGILL
OPIX -- Thread cleanup done for abended pink process 23406
...Starting service thread: CursorServer.
# Starting service thread: WindowServer.
Dumping core after receiving signal SIGILL

ERROR: ::pthread_mutex_destroy ( (pthread_mutex_t*) fSemaphore ) was EBUSY in fi
le Synchronization.C, line 2353
PThreadCheck Stack Crawl
Stack crawl:
        StackCrawlDumpCurrentStack
        PThreadCheck(int,const char*,const char*,int)
        TMonitorLock::~TMonitorLock()
        __sterm_x_2fshemp4_2fTaligent_2fPortable_2fAES_2fAlbert_2fMain_2fScreens
_2fDisplay_2fDisplayLib_5fAllCSources_2eC()
        __C_runtime_termination()
        exit
        cma__core
        cma__sig_deliver
        cma___sig_sync_term
OPIX -- Thread cleanup done for abended pink process 23415
NameServer: removing registered task (5,23415) named "ColorTableServer"
...Thread (2, 22906) "HackInputDeviceHandler" didn't catch an exception, terminate called.
Processs 22906 about to die.
You may run this under a debugger. It should break in raise(),
allowing you to see the calling sequence to the routine that
threw the exception.
Dumping core after receiving signal SIGIOT
OPIX -- Thread cleanup done for abended pink process 22906

Eventually the CommonPoint application system just hangs.

If you are running the CommonPoint application system remotely (by setting your DISPLAY environment variable), make sure that the specified display exists and can be opened.
If you are not running the CommonPoint application system remotely, this symptom is probably caused by incorrect permissions somewhere within $TaligentRoot.

Shared libraries are loaded into a system-wide segment and are intended to be shared. For this procedure to work, shared libraries must be world-readable, otherwise they can't be accessed by other processes.

To ensure proper permissions when installing the system, set your default umask to 022. To ensure proper permissions after the system has been installed, go to $TaligentRoot and execute chmod -R o+r * .

[Contents] [Previous] [Next]

Click the icon to mail questions or corrections about this material to Taligent personnel.

Generated with WebMaker