[Samba] Weird issue - STATUS_DISK_FULL
nkadel at gmail.com
Sat Mar 28 11:52:20 MDT 2015
On Sat, Mar 28, 2015 at 7:50 AM, Ciaran Scolard <Ciaran at phonovation.com> wrote:
> uname -a
> Linux localhost.localdomain 3.19.1-1.el7.elrepo.x86_64 #1 SMP Sat Mar 7 13:55:36 EST 2015 x86_64 x86_64 x86_64 GNU/Linux
> rpm -qa | grep -i samb
> Local I/O appears to be unaffected.
> btrfs balance start was run prior to the copy attempt.
Check the usage of inodes, and use 'lsof' to check for deleted files
being held oopen. One of the great amusements of my life is when
people do silly things like create complex, bulky hardlinked arrays of
deep backup trees with "cp -al" and rsync commands that preserve
unchanged files as hard links, and forget that every directory in
every hardlinked directory tree is, itself, a distinct inode. Even if
the files are hardlinked and take up no space, you can run yourself
out of inodes with tools like "rsnapshot" if you don't pay a bit of
attention tho this.
More information about the samba