Based off what you wrote - and the fact that I’m massively sleep-deprived - it all makes sense. The issue you describe and the fix applied are akin to what we see in the database world, where users complain about queries being slow or unresponsive after trying to force-kill. Only for us to find out, that they submitted queries with a COMMIT after the whole 10mil record transaction, which clearly the DB can handle, but it will take a significant amount of time to rollback vs if the COMMITs are broken up and submitted more frequent. Basically chunking up the data into more manageable pieces as to not saturate the db threads, not to mention the underlying REDO and transaction log files too. So hope this was truly a long-term fix vs just a short-term one. Either way, great write up.( Also, you may want to invest in some 2.5gb networking for later, not that 1GbE isn’t enough, but just more pipeline is great, although I don’t know how much upgradeability your Synology will have in that department, so YMMV)