Since you must keep the system running 24/7, then upgrading the firmware will probably have to be scheduled. I wonder if the RSLinx Enterprise patches do any harm if those are performed first?
WIth that on the backburner, how about designating one of the three programmers to have ownership of all edits that involve adding/deleting tags, and schedule your online edit accept, test, and edits on mutually exclusive intervals to give the CPU time to handle that extra load. I think incremental online edits are more harmful to CPU rsources than larger planned ones. Can you perhaps develop program changes offline, then drag and drop new programs into a running PAC (with a unique strategized name) and then schedule them while running to swap in place of other routines? Would that help with the fragmentation of CPU resources?
I dunno, just a thought. But if you take the system down or it faults, I would at least take that opportunity by having a freshly compiled latest and greatest program ready to download, and burn to CF card. That should help clean up fragmented memory due to online editing and should only take a couple of minutes tops.
It could open up the dilemna of capturing real live running process values, but if the system crashes don't you lose that anyway?
