It is not possible to start a ploop-based container on Parallels Cloud Server 6.0 with the following error:
~# vzctl start 100 add delta dev=/dev/ploop1 img=/vz/private/100/root.hdd/root.hds (rw) Mounting /dev/ploop1p1 at /vz/root/100 fstype=ext4 data='balloon_ino=12,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,pfcache_csum,pfcache=/vz/pfcache' Container is mounted Starting the Container ... Setting permissions 20002 dev 0x7d00 Adding offline management to Container(1): 4643 8443 4649 Adding IP addresses: Setting permissions 60010 dev 0xb611 Old quota version detected Quota utilities version 3.13. Quota should be > 3.16 Error setting up ugid quota Unmounting file system at /vz/root/100 Unmounting device /dev/ploop1 Container is unmounted Failed to start the Container
As of Parallels Cloud Server 6.0, containers are using new layout and filesystem inside containers is Ext4.
Some old distributions like CentOS 5, RHEL 5, Debian 5 have the outdated quota tool (
<3.16) which does not recognize Ext4 filesystem, thus if the container has second level quota limit (quotaugidlimit) set, then the
quotacheck command invocation fails.
The issue is permanently fixed in Parallels Cloud Server 6 update 3:
Article #2509925 Parallels Cloud Server 6.0 Update 3 Hotfix 1 (6.0.0-1323)
After installing the update disable journal quota for already existing containers which have the problem:
~# vzctl set CTID --jquota no --save
For older versions of the product it is possible to disable second-level quota inside the containers:
~# prlctl set CTID --quotaugidlimit 0