Recently I started evaluating if the Duplicacy CLI (linux binary) can be an option to backup my Synology NAS to Backblaze B2 as the Cloud Sync Synology implementation for B2 has so many flaws (mainly it doesn’t encrypt file names, it is not block based in any way so it reuploads whole files even if 1KB change in a 100GB file and it is wasteful in other ways with B2 space).
So first of all I was happy to see the linux binary just work on my NAS when I SSH to it. So it is super easy to set up a manual scheduled task to backup the NAS with Duplicacy which is great.
The initial backup (~1,5 TB of files with very varying size) took about 5 days which is exactly what was to expect when my uploads speed was used at nearly 100%. So no bottleneck there and good job so far.
However now the process for follow-up backups takes about 10 to 15 hours (2 tests so far and the third currently running) where it is noteworthy that there is very little upload (about 1 to 10 GB which should be uploaded in under one hour) and also nearly no disk IOPS, read/write or utilization. The CPU is constant at 25% for the duplicacy task. So as my Synology has an Intel Quad Core my hunch is it is using only one core and the process could maybe be accelerated a lot when it would use all or at least 3 cores. (Needless to say I’m using Duplicacy with encryption.)
So my questions now are:
- Can I achieve somehow faster follow-up backups / use more CPU?
- Is this in general a good setup or should I look for other solutions (which I don’t like as I like the principles of Duplicacy very much! ;-))