Hacker News Reader

Comment by dawidpotocki | Hacker News Reader

I have personally moved to gdu (https://github.com/dundee/gdu) as I found it to be way faster than ncdu (I haven't checked the fork but I don't expect it to improve that much).

1 minute vs 1 second for my entire / which includes mounts to NFS shares (used 1.4 TiB on host SSD, 14 TiB on NFS HDDs and 3 TiB on NFS SSD).

View on HN

Could you do me a favor and test with

   ncdu -t32

this uses 32 threads and should be much faster. I think only reason this isn't default is to be a friendly tool on your host.

It's been going for 5 minutes already… I can't be bothered to let it run longer.

Seems to be wasting a lot of time scanning `/run/user/1000/doc` from Flatpak.

If you use `-x` you'll stay on the same filesystem (so not needlessly check tmpfs or other special filesystems)

So that will exclude other partitions mounted which is e.g. completely useless on my storage server. This and having to remember to use flags or I could just use `gdu` which just works.

Thanks for the tip! Stopped me from installing yet another thing.

Reply on news.ycombinator.com