On Sun, Apr 06, 2008 at 06:09:39PM +0200, Huszár Viktor Dénes wrote:
> Hi guys, the earlier problem came out again:
>
> Here are the logs:
>
> [root]~# df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/md0 42G 15G 26G 37% /
> tmpfs 7.9G 0 7.9G 0% /lib/init/rw
> udev 10M 76K 10M 1% /dev
> tmpfs 7.9G 2.2M 7.9G 1% /dev/shm
> /dev/md1 420G 31G 389G 8% /var/www/users
> /dev/mapper/a-a 17T 17T 51G 100% /var/www/users/sms
So you effectively have a full filesystem. Inode allocation
may be impossible as the filesystem approaches full as there
is not enough contiguous free space to allocate them in.
Also, you are probably using the inode32 allocator (i.e. no inode64
mount option), which means inodes can only be placed in the first
1TB of the filesystem, and that means those AGs are probably out of
space.
Given that you have 523GB AGs, that means inode will probably only be
put in AG 0, ad that's probably out of space. What does:
# xfs_db -r -c 'freesp -s -a 0' /dev/mapper/a-a
give you?
FWIW, This problem will come and go as you unlink files and change
the pattern of freespace...
Cheers,
Dave.
--
Dave Chinner
Principal Engineer
SGI Australian Software Group
|