Home / server / Questions / 723909
In Process
A T
Asked: 2015-09-22 19:25:26 +0800 CST

>11GB free space, but it says 100% disk usage?

  • 772
$ df -i /mnt/large_linux/
Filesystem       Inodes   IUsed    IFree IUse% Mounted on
/dev/sda1      29376512 1035086 28341426    4% /mnt/large_linux
$ df /mnt/large_linux/
Filesystem     1K-blocks      Used Available Use% Mounted on
/dev/sda1      462509952 451315844         0 100% /mnt/large_linux
$ sudo dumpe2fs /dev/sda1 | grep -i reserved
dumpe2fs 1.42.12 (29-Aug-2014)
Reserved block count:     5875200
Reserved GDT blocks:      995
Reserved blocks uid:      0 (user root)
Reserved blocks gid:      0 (group root)
  Reserved GDT blocks at 30-1024
  Reserved GDT blocks at 32798-33792
  Reserved GDT blocks at 98334-99328
  Reserved GDT blocks at 163870-164864
  Reserved GDT blocks at 229406-230400
  Reserved GDT blocks at 294942-295936
  Reserved GDT blocks at 819230-820224
  Reserved GDT blocks at 884766-885760
  Reserved GDT blocks at 1605662-1606656
  Reserved GDT blocks at 2654238-2655232
  Reserved GDT blocks at 4096030-4097024
  Reserved GDT blocks at 7962654-7963648
  Reserved GDT blocks at 11239454-11240448
  Reserved GDT blocks at 20480030-20481024
  Reserved GDT blocks at 23887902-23888896
  Reserved GDT blocks at 71663646-71664640
  Reserved GDT blocks at 78675998-78676992
  Reserved GDT blocks at 102400030-102401024

However I have plenty of free space:

$ pydf -k
Filesystem      Size      Used    Avail Use%         Mounted on                
[omitted]
/dev/sda1  462509952 451315844 11177724 97.6 [#####] /mnt/large_linux

GParted

I've also tried running sudo tune2fs -m 0 /dev/sda1, to no avail.

linux

1 Answers

  1. I don't see in the info you provided any reason for which you'd think you have 11G left (other than the difference in G between Size and Used + Available, but it's incorrect to assume that such difference means Available, especially when Available is clearly stated as 0). On all extX filesystems Used + Available < Size due to reserved blocks and maybe other reasons.

    There is no discrepancy between the df and pydf reports on the respective columns (as you questions appears to suggest) if you take into account that 1G = 1024 * 1024 K:

    • int(462509952/1024/1024) = 441
    • int(451315844/1024/1024) = 430

    Plus both df and pydf indicate 0 on the available column.

    • 0