Can't enable user quota after upgrade from 2.3 to 3.0

[geomaster]

CT config

ONBOOT="yes"

PHYSPAGES="0:1572864"
SWAPPAGES="0:1572864"
KMEMSIZE="2927624192:3221225472"
DCACHESIZE="1463812096:1610612736"
LOCKEDPAGES="786432"
PRIVVMPAGES="unlimited"
SHMPAGES="unlimited"
NUMPROC="unlimited"
VMGUARPAGES="0:unlimited"
OOMGUARPAGES="0:unlimited"
NUMTCPSOCK="unlimited"
NUMFLOCK="unlimited"
NUMPTY="unlimited"
NUMSIGINFO="unlimited"
TCPSNDBUF="unlimited"
TCPRCVBUF="unlimited"
OTHERSOCKBUF="unlimited"
DGRAMRCVBUF="unlimited"
NUMOTHERSOCK="unlimited"
NUMFILE="unlimited"
NUMIPTENT="unlimited"

# Disk quota parameters (in form of softlimit:hardlimit)
DISKSPACE="30G:33G"
DISKINODES="6000000:6600000"
DEVNODES="sdz2:qrw sdz1:rw"
QUOTAUGIDLIMIT="2000"
QUOTATIME="600"

# CPU fair scheduler parameter
CPUUNITS="3000"
CPUS="4"
HOSTNAME="fs.kmz"
SEARCHDOMAIN="kmz"
NAMESERVER="192.168.10.1"
IP_ADDRESS="192.168.10.5"
VE_ROOT="/var/lib/vz/root/$VEID"
VE_PRIVATE="/stor/ct/private/10005"
OSTEMPLATE="centos-6-standard_6.0-1_i386.tar.gz"

In-container fstab

none    /dev/pts        devpts  rw      0       0
/dev/sdz1       /Group  ext3    defaults        0       1
/dev/sdz2      /home   ext3    usrquota,defaults       0       1

In proxmox ve 2.3 quotas on /home mount point work fine. After upgrade /home not mounted, in dmesg (on host and on CT) i see this lines:

VE10005: no allowed mount options found for device 8:34

Without "usrquota" parameter /home mount ok.

 

[geomaster]

Anybody help me?

Отправлено с моего HTC Wildfire через Tapatalk

 

[geomaster]

Run on host

mount -o usrquota,remount -n /dev/sdz2 /var/lib/vz/root/10005/home

And voila! User quota in container is on!

]# cat /etc/mtab
/dev/sdz2 /home ext3 rw,relatime,errors=continue,barrier=0,data=ordered,usrquota 0 0

]#quotaon -p /home
group quota on /home (/dev/sdz2) is off
user quota on /home (/dev/sdz2) is on

I put this magic command into 10005.mount script and pray that it worked on restart...

10005.mount:

#!/bin/bash
. /etc/vz/vz.conf
. ${VE_CONFFILE}
SRC=/dev/sdz2
DST=/home
if [ `mount | grep ${SRC} | grep usrquota | wc -l` -eq 0 ]; then
  mount -o usrquota,remount -n ${SRC} ${VE_ROOT}${DST}
fi

 

[geomaster]

Sad... The action script $VE_ID.mount run BEFORE VE is running. And devnodes is not exists into container. The $VE_ID.start script run AFTER VE is running, but the start and stop scripts are performed in the VE context.

 

[geomaster]

I found some workaround and my VE_ID.mount looks like this:

#!/bin/bash

echo "#!/bin/bash
sleep 5

_sleep() {
status=(\`/usr/sbin/vzctl status ${VEID}\`)
if [ \${status[4]} == \"running\" ] ; then
        . /etc/vz/vz.conf
        . /etc/vz/conf/${VEID}.conf
        SRC=/dev/sdz2
        DST=/home
        if [ \`mount | grep \${SRC} | grep usrquota | wc -l\` -eq 0 ]; then
                mount -o usrquota,remount -n \${SRC} \${VE_ROOT}\${DST}
        fi
        rm -f /etc/vz/${VEID}.sh
        exit 0
fi
sleep 1
_sleep
}

_sleep" > /etc/vz/${VEID}.sh
chmod 755 /etc/vz/${VEID}.sh

. /etc/vz/${VEID}.sh &
disown
exit 0

It run and re-mount devnode with usrquota. But I can't test this script with migration.