Hi,
I have 3 unattended bot which need to run on a file having 100000 records in parallel. Should I need to add all the records in a queue? Is there any step by step reference available? How to configure the bots in Orchestrator so that they can run parallel.
You dont need to add to a queue if you dont want to, i assume you have no process developed yet right? If this is something you are running only one time, maybe you just want to give each of your robots their own data to process, but i recommend that you use the queues for better management of the results you are getting, you will see examples of this kind of process in the Academy. Your robots should be already configured in Orchestrator, what you will need to process the queue will only be a job execution selecting your 3 robots, the rest will be on the process you develop.
This is a very high number of transaction items and will take some time to load into the queue if you choose to use one.
However, you could create a dispatcher to load the items, and have the queue trigger bots to process the items, while the dispatcher is still loading new items in the meantime. Then you won’t have to worry about the time it takes the dispatcher to load the items, since your performers will be running simultaneously.
Thanks for the response. One more doubt how the data (rows) will be distributed among 3 robots? It will be a continuous process as we will receive data file everyday.
The process will query the data from the queue and get the next available transaction and process it, since you will have 3 robots running the same process, each will get a different transaction…
Look into this for bulk uploading of transactions:
You could use the Bulk Add activity (@bcorrea correctly pointed out) , but 100000 records could be painful to monitor and review, though. Another approach would be to add the file path as the transaction rather than the 100000 records in the file… or organize them differently.
To answer how to get multiple robots to manage the data at the same time, you will need to surround the Write/Append range steps with a Retry scope. And, set the retries up so it will keep trying to update the file for let’s say 1-2 minutes. You just need to make sure that the method used to write the updated data to file errors when the file is in use. I have been able to use this approach with up to 10 robots which I have tested.
Regards.