[Ltsp-discuss] Filesystem /cow: no space left on device

[Rainer Stumbaum]

Hi,
we have some LTSP clients with a high uptime that get a filesystem full on /cow:

***@controlnode01.dc1:~# cd /cow
***@controlnode01.dc1:/cow# df -h .
Filesystem Size Used Avail Use% Mounted on
tmpfs 502M 456M 47M 91% /cow
***@controlnode01.dc1:/cow# mount | grep aufs
aufs on / type aufs (rw,relatime,si=8e2f2afaba323e27)
***@controlnode01.dc1:/cow# mount | grep cow
tmpfs on /cow type tmpfs (rw,relatime,mode=755)
***@controlnode01.dc1:/cow# du -h -d 1 .
2.0M ./usr
28K ./sbin
132K ./etc
1.7M ./tmp
38M ./var
0 ./.wh..wh.orph
7.3M ./.wh..wh.plnk
49M .
If we delete files the du goes down but df stays the same.
According to https://www.mail-archive.com/aufs-***@lists.sourceforge.net/msg04321.html
LTSP should possibly use some extra mount options for the aufs filesystem.
trunc_xino
Has this been tried and where do we modify that?
Thanks
Rainer

P.S.: Is the bug reporting system on https://bugs.launchpad.net/ltsp actually used? I reported some bugs and patches there and did note yet receive a response.

[Vagrant Cascadian]

Post by Rainer Stumbaum
P.S.: Is the bug reporting system on https://bugs.launchpad.net/ltsp
actually used? I reported some bugs and patches there and did note yet
receive a response.

Yes it is used. Sorry you haven't gotten a response...

Unfortunately there aren't many active developers to review bugs or even
patches.

I try to address issues as they ome in if they're obvious, or need more
clarification, but ones that need further exploration tend to sit a bit,
especially if it's specific to another distro (my personal focus is on
Debian, and occasionally Ubuntu).

For better or worse, the #ltsp channel on irc.freenode.net tends to be
the easiest space to sort out issues... but no guarantees.


live well,
vagrant

[Jeff Siddall]

Hi,

I have also seen "no space left" but have never looked into it further
than confirming there is space left on the server side (so this is
something on the client).

Jeff