Syncing 5 VPS to one storage backend with rclone. I got these cheap Contabo and OVHcloud boxes scattered everywhere and I got tired of SSHing into each one to grab logs. Made a cron setup that runs every 6 hours, pushes everything to a Hetzner storage bucket with --bwlimit 2M so I dont get nullrouted.
We have seen both approaches in production environments. Rclone with bandwidth limiting offers predictability for metered links, while mergerfs provides a unified namespace without the scheduling complexity.
For five nodes, we typically recommend testing both under your actual workload rather than assuming either is superior. Your I/O patterns matter more than the tool.
Contabo nullrouted me once at 1.8M even with bwlimit. Their traffic shaping is aggressive. I switched to OVHcloud for the heavy sync jobs, no issues yet.
This is the key question. Mergerfs handles offline nodes by design -- reads fail gracefully, writes to the remaining branches continue. But your backup completeness suffers, and you only notice when you try to restore.
With rclone cron you get explicit failure logs. I know which node missed its window.
