Guys,
I’m kicking off the same service in the scheduler 2 times in very close proximity with different values for the input parameters. It appears each job is not getting its own variables. The 2nd job seems to gets very confused and has botched parameters resembling that of the 1st job.

Anyone seen this before and know how to get around it? 

Basically I need the same service to run multiple instances to pick up ftp files.  number of jobs and parameters is DB driven.  So you can turn on a new job and set parameters in the database.   

Anyhow works beautifully until we try to launch more than one job for the same service.

You try calling …flow:clearPipeline to clear the pipeline at the end of the service.

Rob,
There definitely will be time when these are running in parallel.
So i don’t think clear Pipeline is going to help in that situation because that will not be executed until the very end, and there could already be another parallel task running the same service.

That is strange - I would think each scheduled service would run in its own thread with its own set of variables. We had problems with the scheduler before dropping db connections when multiple processes had been kicked off with the scheduler. We created a java service that the scheduler would run which would reconnect to the server and invoke a service using the ‘Context’ class so that each service had it’s own session. The only problem is that you need to obtain the username/password of a user in IS when you use these API calls. Check out the java API under the Developer directory.

Guys,
Beware i found the problem… You should schedule more than one job from the same session dynamically. If you schedule 2 jobs running the same service even at different times there input and run-time variables get corrupted.

I got around this by scheduling a job that runs only to launch other jobs if necessary. It only launches one at a time thus ensuring that we have a new session each time because the master job stops and starts and gets a new session.

Hi, I have a scheduled service that runs every 5 seconds. It works initially but then the next run indicator returns very unpredictable values eg. -13232123 or large numbers, 13131231212. This results in the service not being run at all. My configuration for the scheduled service is: Run as Administrator, Persist after restart, Repeat from end of invocation.
Can anyone help or explain?

Hi,

It is generally a know problem that the schedular takes large or negative numbers, when it at that particular point of time not able to compute when it would run again precisely or when the current execution would end.

How ever this should not hinder it from executing appropriately as required/configured by you. The large/negative number is an internal depiction.

On careful observation you should be able to see the large number change once the existing execution comes to an end and the scheduler can precisely know the next execution time.

Just to cross check try to observe Uncheking the, “Repeat from end of invocation”.

Hope that gives a good insite into the problem.

Thanks, I’ll try that. I see that WM has a fix for this: IS_4-6_SP2_Fix179. But I can’t seem to find it. Do you know anything about this?

Guess you have to contact webMethods support for this, no big clue on that.