EDX Command Line Fails but QMC Task Continues to Run When Multiple Docs Are Reloading in Parallel

[robsilcock]

Hi.

For the last 2-3 years my project has been using the EDX command line to trigger document reloads configured in our QMC.  We use an external scheduler (Control-M) to trigger batch files on our QlikView servers and these bacth files in-turn trigger the EDX component and pass to it the task name of the document that we want to reload. (Nothing complicated, quite simple really.)

The EDX executable and associated config file are saved in a central location.  Every batch file that is triggered uses this single central instance of the EDX component.

Up until recently our batch has had a serial structure where one document reload follows on after the successful completion of another.

We have recently introduced a new branch to the batch that reloads a document in parallel to the existing document reloads.

We are now seeing regular batch failures on an almost daily basis where EDX fails but the documents reloading in the QMC run through to a successful completion.

The error details passed back to Control-M seem to suggest that EDX is losing its thread with the active task when it polls the QMC to check the task's status.

Does anyone on the community have experience of this issue?

Can anyone recommend any best practise regarding EDX command line?  For example, is it ok to have multiple documents running in parallel that were trigged from a single central instance of the EDX component?

[prabhu0505]

Unless there is starvation for cores (resources) to execute, you can do polling kind of thing for any long time.

Assuming Multi-box environment and QDS is in separate box:

(1) 8 cores in server, can be configured for ~5 concurrent tasks

(2) for 12 cores server ~8 concurrent tasks should be fine.

So depends on the environment configuration.

[robsilcock]

Hi Saravana, thanks for your reply - much apprecaited!

Can you please help by confirming if the conclusions I am extracting from your reply are correct?

(1) The issue that we are seeing is more than likely resource related (i.e. cores and/or RAM) on the host server where the QMC is hosted that is implementing the document reloads?

(2) It is therefore perfectly acceptable to have just a single central instance of the EDX command line component and this can be called by multiple doc reloads that are running in parallel at the same?

Kind regards .... Rob