I routinely have to run file system scans on a giant NFS filer, and even without a hiccup, out of a 100M stat or read calls, ten or so will just never finish. In Go, I have to wrap the call with a channel thing and a time out and hope I don’t run out of threads before scanning all 400 M files.