I have a large batch process that interacts in a page hosted by a cloud service. The process is ran by 3 resources, 1 master and 2 slaves--they work out of the same queue of up to 1000 records at a time.
I've noticed that over the course of a couple hours, as the process runs, the browser will occasionally stop responding which doesn't come as a surprise since it's working in a browser for hours at a time. I created a re-spawn routine for the slave resources which closes the browser, initiates the process again and the slaves continue working out of the same queue. The master bot however is programmed to truncate the queue at the beginning of processing under the assumption that it's initiating a new batch UNLESS a rerun environment variable is flagged to true, in which case the Master will behave as a slave and just continue processing out of the existing queue.
My question is, is it possible to change environment variables at runtime? My thought is, for the master bot, if it's determined it needs to re-spawn, the rerun environment variable can be switched to true in order to preserve the existing queue and then it can continue to process records until it completes at which point the environment variable can be switched back to false for the subsequent job, the following day.
Thought? Suggestions?
Thanks
JP