nil semaphore in SstWorker

[Robert Webb]

One of our customers received the exception 'UndefinedObject does not understand wait' in SstWorker>>#workBlock. Any ideas why the semaphore would be nil? This is in VisualAge v6.0.2.

UndefinedObject(Object)>>#error:
self=nil
aString='UndefinedObject does not understand wait'
UndefinedObject(Object)>>#doesNotUnderstand:
self=nil
aMessage=a <Dumped> Message{
selector: #wait
arguments: a <Dumped> Array
}
[] in SstWorker>>#workBlock
self=a <Dumped> SstWorker{
work: nil
name: 'Dispatch worker: %1'
priority: 3
semaphore: nil
process: nil
processNumber: nil
}
current=nil
[] in Process>>#executeBlock:withArguments:
self=a <Dumped> Process{
returnPoint: nil
vmState: 975680
stack: a <Dumped> EsLongArray
priority: 3
state: ##running
queue: a <Dumped> EsQueue
debugBP: nil
switchCount: 7
errorHandler: [] in <Dumped> Block>>when:do:exitWith:retryReturn:
resumeable: true
breakBP: nil
interruptBlocks: nil
attributes: a <Dumped> LookupTable
terminateBlock: [] in <Dumped> Process>>executeBlock:withArguments:
processName: 'Dispatch worker: 45'
blockingObject: nil
}
aBlock=[] in <Dumped> SstWorker>>workBlock{
methodContext: <Dumped> SstWorker>>workBlock
parent: a <Dumped> SstWorker
startPC: 11
argsAndTemps: 32903
nativeStartPC: 67
}
args=a <Dumped> Array{
}
Process>>#executeBlock:withArguments:
self=a <Dumped> Process{
returnPoint: nil
vmState: 975680
stack: a <Dumped> EsLongArray
priority: 3
state: ##running
queue: a <Dumped> EsQueue
debugBP: nil
switchCount: 7
errorHandler: [] in <Dumped> Block>>when:do:exitWith:retryReturn:
resumeable: true
breakBP: nil
interruptBlocks: nil
attributes: a <Dumped> LookupTable
terminateBlock: [] in <Dumped> Process>>executeBlock:withArguments:
processName: 'Dispatch worker: 45'
blockingObject: nil
}
aBlock=[] in <Dumped> SstWorker>>workBlock{
methodContext: <Dumped> SstWorker>>workBlock
parent: a <Dumped> SstWorker
startPC: 11
argsAndTemps: 32903
nativeStartPC: 67
}
args=a <Dumped> Array{
}
Process>>#newProcessOn:stackSize:withArguments:named:
self=a <Dumped> Process{
returnPoint: nil
vmState: 975680
stack: a <Dumped> EsLongArray
priority: 3
state: ##running
queue: a <Dumped> EsQueue
debugBP: nil
switchCount: 7
errorHandler: [] in <Dumped> Block>>when:do:exitWith:retryReturn:
resumeable: true
breakBP: nil
interruptBlocks: nil
attributes: a <Dumped> LookupTable
terminateBlock: [] in <Dumped> Process>>executeBlock:withArguments:
processName: 'Dispatch worker: 45'
blockingObject: nil
}
aBlock=[] in <Dumped> SstWorker>>workBlock{
methodContext: <Dumped> SstWorker>>workBlock
parent: a <Dumped> SstWorker
startPC: 11
argsAndTemps: 32903
nativeStartPC: 67
}
stackSize=1024
args=a <Dumped> Array{
}
procName='45'
==============<STACK TRACE ENDS>=============

[wembley]

Robert -

From the information you sent, it looks like SST was trying to shutdown (which explains why all the variables in the SstWorker were nil) and something got in and tried to use one of the cleaned-up SstWorkers.

Can you tell me anything more about what is going on in the system when this error happens? Do you see any other errors around the same time?

[Robert Webb]

The system was creating a new dispatch worker, presumably to respond to requests coming in through RMI. The fact that this one was labeled #45 indicates that a spike in activity was occuring, since we normally only see 4 or 5 dispatch workers in use.