On Mon, May 29, 2006 at 01:54:54AM +0200, diego torres wrote:
> hi there!
> I'm running some vservers, and one of them with context quotas
> enabled. After copying all the servers to a new disk, the one with
> context quotas, when using df -h doesn't report the used space
> correctly. It reports a lot less. Looks like old files hasn't been
> tagged correctly. With new files there is no problem, and space
> decreases and increses correctly.
> I'm not using hashify/vhashify.
> Because of this, I have used chxid with no errors:
> /# chxid -c services -R /home/vservers/services/
> but it doesn't solve the problem.
> This is the used space:
> /# du -hs
> 3.9G .
> and this is the report of df.
> /# df -h
> Filesystem Size Used Avail Use% Mounted on
> /dev/hdv1 5.0G 12M 4.8G 1% /
> none 1.0G 0 1.0G 0% /tmp
> Obviously "Used" should be 3.9G or more, butn ever less. In the rest
> of vservers, du and df output makes sense. 5Gb is the quota limit, and
> /tmp of 1Gb has also been configured on purpose.
> Making the long story short, it looks like chxid isn't working (that's
> the best I have worked out)
you probably want to recalculate the disk space
for that guest, IIRC this can be forced by removing
the dlimit information in the config tree and
restarting the guest, but you could also just set
the correct values with vdlimit
> Some version info:
> kernel + patches: 126.96.36.199-vs2.1.1-rc21
> util-vserver: 0.30.210-10
> debian testing
> The problem with this is that 4.8Gb are available, instead of 2.2Gb
> that should be the right thing. Any idea in how to force the system to
> realize the current space?
> Vserver mailing list
Vserver mailing list
Received on Tue May 30 18:16:12 2006