cristihcd

Adding to this. Using kopia server on Windows (webui opened) and 1 kopia client (via webui) on debian. Trying to snapshot 107GB folder with many small files (350000) on debian...

I tested this again with kopia server running on debian linux. On a virtual machine with 32GB RAM, kopia server memory usage slowly goes up to maximum and then is...

I built and tested again with the zap logger branch (setup [here](https://github.com/kopia/kopia/issues/1233#issuecomment-912345342) - 400k small files) and memory usage for kopia server went up to ~60GB. Maybe pprof would help...

It seems to occur only with zstd-best-compression. I also tested with: no compression/zstd-fastest (about the same memory usage -- ~1GB). With zstd-best, memory usage rises quickly to 12GB, then 24GB....

Looks like a welcome feature. I would suggest --skip-existing and --delete-extra flags.

Currently I am using `robocopy /mir` over an on-demand vpn connection to delete extra files. The restored files are over 400GB, so deleting before restore is not a good option....

Here is a comparison (benchmark) based on https://github.com/gilbertchen/benchmarking ### Results (linux kernel): > Here are the elapsed real times (in seconds) as reported by the time command, with the user...

Additional comparison with Duplicacy which I used previously for a Postgresql database. Repo converted from Duplicacy. - Original (source) size: 282GB - Duplicacy repo: 38 snapshots, 546GB, 128000 files -...

testing the following acl with headscale v0.22.3, tailscale v1.46.0 (linux), it seems to work: ``` { "acls": [ { "action": "accept", "src": ["user1"], "dst": ["user1:*"] }, { "action": "accept", "src":...

I think opt-out should be the default, otherwise too little data will be collected. Also see https://data.syncthing.net/ for an example Settings:  Preview: