[Errno 28] No space left on device

jbn

I'm getting

[Errno 28] No space left on device<br>

error messages when trying to copy to my ZyXel NAS540. To me, it looks like there is space available,

~ $ df
Filesystem           1K-blocks      Used Available Use% Mounted on
ubi7:ubi_rootfs2         92892     51160     36952  58% /firmware/mnt/nand
/dev/md0               1966352    192368   1674096  10% /firmware/mnt/sysdisk
/dev/loop0              151719    135159     16560  89% /ram_bin
/dev/loop0              151719    135159     16560  89% /usr
/dev/loop0              151719    135159     16560  89% /lib/security
/dev/loop0              151719    135159     16560  89% /lib/modules
/dev/loop0              151719    135159     16560  89% /lib/locale
/dev/ram0                 5120         4      5116   0% /tmp/tmpfs
/dev/ram0                 5120         4      5116   0% /usr/local/etc
ubi3:ubi_config           2408        96      2156   4% /etc/zyxel
/dev/md2             11616404496 5495211768 6121192728  47% /i-data/ad82b6f7
/dev/md2             11616404496 5495211768 6121192728  47% /usr/local/apache/htdocs/desktop,/pkg

and inodes available,

~ $ df -i
Filesystem              Inodes      Used Available Use% Mounted on
ubi7:ubi_rootfs2             0         0         0   0% /firmware/mnt/nand
/dev/md0                124928        12    124916   0% /firmware/mnt/sysdisk
/dev/loop0               39200      6983     32217  18% /ram_bin
/dev/loop0               39200      6983     32217  18% /usr
/dev/loop0               39200      6983     32217  18% /lib/security
/dev/loop0               39200      6983     32217  18% /lib/modules
/dev/loop0               39200      6983     32217  18% /lib/locale
/dev/ram0               126455         8    126447   0% /tmp/tmpfs
/dev/ram0               126455         8    126447   0% /usr/local/etc
ubi3:ubi_config              0         0         0   0% /etc/zyxel
/dev/md2             365895680    521745 365373935   0% /i-data/ad82b6f7
/dev/md2             365895680    521745 365373935   0% /usr/local/apache/htdocs/desktop,/pkg

I suspect it's an i-node problem. I have a folder with several hundred folders with 1000 files in each of them. But, I can't figure out what's wrong and I've never done any RAID admin.

#NAS_January

Mijzelf

Could it be a user quota problem?

I've never done any RAID admin.

The raid array is a blockdevice like every other. The filesystem doesn't 'know' it's on a raid array. Error 28 is a filesystem error, and so it has nothing to do with raid.

How are you accessing the volume? 

jbn

Thanks for your time, Mijzelf!

I'm accessing via SAMBA, mounted as an OSX volume. I've replicated the problem via a python script (package pysmb) using different user credentials. Both users have unlimited quotas according to the ZyXEL web user interface.

The curious thing : the problem seems to be localized to one sub-tree. I just successfully transferred 500mb to a folder on the same share without problem. Afterwards, I tried to copy the exact same file that just successfully transfered to the failing subfolder -- failed immediately with ERRNO 28.

Mijzelf

Does that folder name/path contain any strange character? 

That error 28 is a [i]local[/i] error on your client. So for some reason the samba client on your Mac spawns this error. Of course it [i]can[/i] be a pass from the server.

jbn

No, it does not. In fact, every character in the misbehaving subtree is in [ABCDEF0123456789], as the file names are hashes of the contents.

That local error exists on both the OSX SAMBA client and the python package one. Any advice on how to experiment with that share via SSH on the NAS? You pointed out it's just like any block device, but I'm oh-so-nervous to manipulate it directly since it's may backup and I don't have room to transfer it to another device while experimenting... 

(Thanks for continuing to try helping me. I'm at a loss. It seems like an odd problem.)

Mijzelf

Any advice on how to experiment with that share via SSH on the NAS?

You can try to copy that file from the other subtree to the misbehaving subtree, to see if that rises an error.
The subtree can be found in /i-data/ad82b6f7/<sharename>
If that also fails, have a look in the kernel log buffer, using dmesg.
If that doesn't fail it should be nice to have a look at the samba logs, but unfortunately they are disabled on the NAS.

 I'm oh-so-nervous to manipulate it directly since it's may backup 

You are aware that it's not a backup if it's the only copy you have? A backup implies that you have another copy of the data. You should have if the data is valuable. Your NAS will fail some day. Just like any storage medium.