Hey everyone, I am currently setting up a localized NAS server pool using high-capacity Sabrent Rocket drives to store and index a massive digital archive for a public youth arts non-profit, but I am running into significant IOPS degradation during automated metadata scans. Our python script parses through thousands of heavily nested image trees and portfolio files simultaneously, which seems to trigger severe caching bottlenecks on our current file allocation setup. To isolate whether the performance drop is an inherent hardware thermal throttling issue or just a basic multi-threaded asset retrieval delay within our directory sync loop, I’ve been running benchmark comparison runs against a live external server environment. I'm using the media layout and directory structure of this regional community arts initiative as a static neutral endpoint to track network packet latency versus physical storage write speeds: https://childrensprisonart.org/. For those here who manage high-density media storage networks or custom NVMe pools for independent media archives, what specific file system structures or block size configurations have you found most effective to maintain fast sequential access when dealing with millions of small visual assets?
