Max Executed Steps Exceeded

[Phil Bradley]

Hello All. I ran into a situation where a caller got aborted due to:

com.cisco.wfframework.obj.WFMaxExecutedStepsExceededException: No. of executed steps: 1000

I have read other posts related to this and understand that this is due to the number of steps in a script going over the default 1000 steps. However looking at the call start and end time I am confused to why this happened. I have a queue loop setup in my script that delays every 30 seconds and the total duration of the call was around 19 minutes. So one of the calls that got aborted went through the loop around 38 times. Do annotations in a script count as a step? If so then this may be part of the reason we reached the greater than 1000 steps. Also, what issues can occur if I set the system parameter to greater than 1000 steps? I have attached my script and thanks for any input.

[Gajanan Pande]

Phil,

Increasing No. of executed steps more than 1000, is not recommended as 1000 steps is more than enough to wait for. Increasing the count will definitely impact performance & might lead to deployment being unsupported. I'll try to locate a Cisco document which has supportability guidance on this but from my experience, this is what it is.

Pls rate the post if it helps.

GP.

[Phil Bradley]

Hello Gajanan,

Thanks for the reply. The problen with leaving this at 1000 is that I am only getting a maximium of around 25 minutes in queue. I did find out that annotations where included in the step count so I removed these and that helps some. I had 7 annotations in the queue loop that are now removed. I have also increased the max executed steps to 2000 on advise from TAC. The only thing I was warned about is this is what keeps the "infinite loops" from running away with the processor and I am fully aware of this. This increased the time a caller has to remain in the queue to over an hour. My theory is that if a customer wants to hang on for an extended amount of time then the system should support it. We are small and only have a T1 coming in so the load is next to nothing on the server. I will keep a close watch on the server though and make sure that performance does not degrade by doing this. Again thanks for the reply.

[Aaron Harrison · Accepted solution]

Hi Phil

I quite regularly increase the step count on my deployments. Typically this is for management scripts that perform a lot of xml read/write and other validation and may potentially loop around complex stuff many times. As you say, it's there as an infinite loop break - no infinite loop, no problem.

You can keep things like your labels and annotations by setting them on the step customiser for an existing step rather than as individual steps (see the three little tabs on the left of each step customiser) to keep the step count down as well.

Regards

Aaron Harrison

Please rate helpful posts...

[Phil Bradley]

Hi Aaron,

Thanks for the info. I increased the step count and have added the annotations to the actual steps and that has increased the total available queue time to an acceptable value.

Thanks,

Phil