KinMage:src thomas$ sudo mount -o vers=3,intr 172.16.1.129:/fooper /mnt KinMage:src thomas$ ls -la /mnt total 8 drwxrwxrwx 2 root wheel 4096 Feb 13 21:50 . drwxrwxr-t 24 root admin 1224 Feb 13 21:54 ..
Where 172.16.1.129 is snakey, the Linux VM I am using for testing.
And then:
KinMage:mnt thomas$ python punch.py KinMage:mnt thomas$ ls -la p*out -rw-r--r-- 1 thomas staff 1023 Feb 13 21:54 p1023.out -rw-r--r-- 1 thomas staff 1024 Feb 13 21:54 p1024.out -rw-r--r-- 1 thomas staff 1025 Feb 13 21:54 p1025.out -rw-r--r-- 1 thomas staff 10250 Feb 13 21:54 p10250.out -rw-r--r-- 1 thomas staff 102500 Feb 13 21:54 p102500.out -rw-r--r-- 1 thomas staff 64 Feb 13 21:54 p64.out KinMage:mnt thomas$ du -sh p*out 1.0K p1023.out 1.0K p1024.out 1.5K p1025.out 10K p10250.out 100K p102500.out 512B p64.out
So it creates sparse files across NFS!
Well, yes and no. It will only send a block of data across and then the
server OS decides to create the sparse file or not.
Another thing to note is that the size reported is flexible in the sense
that the underlying file system interface determines how much space
is being reported:
[thomas@snakey fooper]$ du -sh p*out 4.0K p1023.out 4.0K p1024.out 4.0K p102500.out 4.0K p10250.out 4.0K p1025.out 4.0K p64.out
Ideally we would like the sizes to match, but since we are pulling a fast one, we get what we see.
