"Crazy" parallel approval step

[Former Member]

Hi all,

I'm using R3 v4.7 and I have a very strage problem in a

workflow to approve purchase requisitions.

I use a parallel step with 7 forks (4 mandatory and 3 optional according to the material groups). In this parallel step I use the deadlines to approve automatically after x days.

In some cases I get the following "crazy" behaviour:

the workflow doesn't continue in this parallel step in the normal way. The green line that highlights the flow in the log has interruptions that I don't understand and some tasks can be sent to any user since some steps are skipped wothout a reason.

I still don't know if there are OSS notes about this issue and I should like to know from you if you have found a similar problem using deadlines. The other approval steps (without deadlines) never had a problem like this.

I'd like to attach a screenshot of the log to be clearer...

Thanks in advance and best regards,

Angelo

[former_member186746 · Accepted solution]

Hi,

I had a similar issue a while back for a condition. The first step after the left branch somehow returned to the right branch, it looked pretty weird in the workflow log.

I then wrote an OSS message and in the end someone fixed it, it was caused by corrupt data in some of the nodes tables, SAP couldn't tell me how this was possible though.

Other times this can happen if you're on DEV with a couple of WF developers, it then helps to refresh all of the buffers or even logging of and logging in again.

Kind regards, Rob Dielemans

[Former Member · Accepted solution]

Hi Rob,

This issue is causing many unpredictable situations in the productive system and I also opened an OSS note.

SAP has requested some data to make their system checks

and the strangest thing is that it happens only sometimes.

I'm using often SWU_OBUF to refresh the buffers, are there other helpful transactions from this point of view?

Thanks, bye.

Angelo

[former_member186746 · Accepted solution]

Well,

most of the times, really strange behaviour wit workflow is somewhat buffer related, or transport related. You could try /$sync and /$tab, but I wouldn't do that on a productive environment, or at least make sure noone is on the systems.

Another thing is, copying the wf template and set up the event linkage for that one.

You can also check if this is caused by something else, I always use sm21 to search for odd behaviour.

Kind regards, Rob Dielemans

[former_member186746 · Accepted solution]

double post

Message was edited by:

Rob Dielemans