We are using TIBCO BW Agent 9.1.4 with some patch files(PBD,PBL and JAR) provided by CA Support to monitor TIBCO BW 5.12
While monitoring Tibco BW (5.12) application, We have got 'Stack Overflow' error in investigator and the BW application went into hanged state without throwing any error. It didn't accept anymore transactions, so application team is asking if 'wily BW Agent' is also using the 'stack size' of the TIBCO application which is assigned by application team.
If stack size is not shared by Agent then what are the other resources apart from GC being used in JVM?
Ying has replied on the case you raised for this problem.
Yes the agent will be creating threads within the TIBCO BW JVM process, so whatever is set for the BW Java Thread Stack Size will need to be increased (I think the BW default is 128Kb)
I would suggest initially doubling the current value and then see how you go.
Will wily agent consume default 128 KB or progressively it will use and if we need to increase stack size from the default value '128 KB' then what should be the size we have to set and how much will be used by WILY agent.
We would like to know what are the other resources apart from GC and stack size being used in JVM?
Currently, we are monitoring TIBCO BW 5.7,5.9 and 5.11 and stack size is default that is 128KB in these BW but we did not face any issue so far but why this issue is occuring for BW 5.12.
We have already set default stack size to '256 KB' and then this stack overflow error occurs.
The stack size parameter will set the maximum value for the stack of each java thread created.
It sounds like your 5.12 environment might have some transactions with deep method calls that are causing some agent threads to grow larger stacks than previously seen.
I would still suggest doubling the stack size to 512KB to see if that will stop the limit being reached.
I think Ying has already updated you on this via the case.
Is this possible to monitor and put alert on stack utilization with the help of Wily?
Any update..Plz suggest.