Question on dynamic allocation with Triggers

Help Orchestrator
1

Hi Commmunity,

Please help me to understand what should happen in the following scenario…

Job A is scheduled to run every 1 hour and is set to run on dynamic allocation of one of 2 server, server 1 and server 2.

Job runs at 10am and is dynalically allocated to run on server 1.

Job takes 3 hours to complete.

Next scheduled run is for 11am.

At 11am would you expect the Job to run now on server 2 meaning Job A is now running on both servers or would you expect the 11am run of Job A to not start and wait for the run on server 1 to complete first?

Hope this is clear…

JOhn.

2

@johnom

if both use same machine template then only one job would be triggered…

if both use different then two jobs can trigger

now if there is only one robot then only one can trigger again…there must be atleast two robots

cheers

3

I think actually it can trigger two jobs, with one robot assuming the robot is configured to allow it to run multiple jobs at the same time on different machines.

There is too little information here to say what will happen, as there are a ton of settings to control this so it works very differently depending on the robot settings, trigger settings etc etc.

4

@Jon_Smith

Is that possible? even I am interested as when we tried the same it never triggered as the user was busy is what we concluded

cheers

5

This is the trigger in question. We think this trigger has kicked off the job on 2 different machines.

See the setting “Execute the Process 1 times”

We thought this might stop the job kicking off for a second time in parallel.

This image shows a configuration screen for setting up a process trigger for downloading invoices, with options for execution target, job priority, runtime type, schedule, timezone, and non-working days restrictions. (Captioned by AI)
This image shows a configuration screen for setting up a process trigger for downloading invoices, with options for execution target, job priority, runtime type, schedule, timezone, and non-working days restrictions. (Captioned by AI)1161×706 73.9 KB

6

@Jon_Smith @Anil_G

I can also confirm that we don’t just think this is happening but can see it happening today.

7

yes the job will be triggered on second server, because (I believe) there is not check in your process for job status.
You can refer the this post, if you dont want to trigger the job on second server if it is already running on first server. It is going to be API call to Orch for get the running job detail.

more info about the activity is here Activities - Get Jobs

8

Yes, this is what I thought. Thanks to everyone for your comments.

9

I believe this setting limits it.

This image displays the "Unattended setup" step in a configuration process, where options include enabling the account for unattended automations, setting credential usage, and an optional choice to run only one job at a time. (Captioned by AI)
This image displays the "Unattended setup" step in a configuration process, where options include enabling the account for unattended automations, setting credential usage, and an optional choice to run only one job at a time. (Captioned by AI)1167×359 34 KB

Aswell as this

The image shows a form configuration for a process named "Mock Performer Trigger," detailing fields such as queue, job priority, runtime type, and job limits in a process automation tool. (Captioned by AI)
The image shows a form configuration for a process named "Mock Performer Trigger," detailing fields such as queue, job priority, runtime type, and job limits in a process automation tool. (Captioned by AI)1196×332 34.1 KB

If you have the robot able to run multiple jobs and two machines (I believe them being connected via the same machine template or not is not a factor) and the trigger set up to allow multiple jobs to be triggered then it will run on both machines.
You can try it out by manually starting a job and allocating it specifically to the same user, same machine template but different host machines.

I actually had the opposite problem where I only wanted something to run on one machine, in sequence, and someone added stuff and made it run in parallel on the same user and mess things up :stuck_out_tongue:

1 Like
10

That additional information does indeed confirm, it will run the second job on the second machine.

My post above shows some ways to limit it.
Now you have shown its a time trigger that also sheds light on the behaviour, there is not control in a time trigger to not start multiple jobs like there is in a queue trigger.