Hi everybody,
we're having problems with the context disk limit facility, which does
not work as expected. Our setup is as follows:
kernel 2.6.12-vs2.0-gentoo-r1
util-vserver-0.30.209
/vservers is a 290gig reiserfs v3 partition, mounted with tagxid.
persistent context tagging is set to uid24/gid24.
Now, suppose i set the following limits:
server12 interface # vdlimit -x 1009 /vservers/
1009 /vservers/
space_used=321272
space_total=3072000
inodes_used=28470
inodes_total=1536000
reserved=0
Now I do in the corresponding virtual server:
vs00009 / # dd if=/dev/zero of=bla bs=1M count=10
10+0 records in
10+0 records out
Everything still looks normal:
server12 interface # vdlimit -x 1009 /vservers/
1009 /vservers/
space_used=331522
space_total=3072000
inodes_used=28471
inodes_total=1536000
reserved=0
But after deleting the file in the vserver, the limit looks like this:
server12 interface # vdlimit -x 1009 /vservers/
1009 /vservers/
space_used=331512
space_total=3072000
inodes_used=28470
inodes_total=1536000
reserved=0
Only 10k were removed from the byte counter although it was an
11meg-file. The "inode" counting seems to work, if it can so be called
on a reiserfs partition.
The whole directory has of course been chxid'ed to the right context id
before the test.
What's going wrong here? Am I overlooking something? Is it a bad idea to
use reiserfs for /vservers? (Considering the code seems to be meant for
ext2/3, with the inodes and the reserved space)
with kind regards,
Cornelius Thiele
_______________________________________________
Vserver mailing list
[email protected]
http://list.linux-vserver.org/mailman/listinfo/vserver
