Article ID: 121453, created on May 3, 2014, last review on May 11, 2014

  • Applies to:
  • Virtuozzo 6.0
  • Virtuozzo hypervisor


Virtual Machine startup process gets stuck in starting state:

[root@node ~]# prlctl list
UUID                                    STATUS       IP_ADDR         NAME
{08cd9db0-425a-649e-955c-5deec2e2c418}  starting

Meanwhile in Virtual Machine's log /var/parallels/ you can see following message:

05-01 20:37:50.931 F /vm:2682:ed06a700/ Sending question = PET_QUESTION_DELAYED_COMPACT_VM_DISKS


Startup gets stuck because Parallels Server asks starter whether he wants to compress VM's disks or not. Normally this question should only be asked when starting VM via Parallels Management Console, when there is an interactive session. Due to bug #PSBM-4391 this question is asked even when VM is started from command line.


This bug will be fixed in future updates of Parallels Cloud Server.

As a workaround you should start VM using Parallels Management Console - starting the VM from interactive session should allow you to cancel disk compressing request. On attempt to start this VM you'll get dialog window asking whether you want to compress disks or not. Completing the dialogue window should unlock the startup process.

If Virtual Machine is still stuck in starting state you can unlock it applying steps outlined below:

  1. While VM is in 'starting' state make sure that last record in the logs reports PET_QUESTION_DELAYED_COMPACT_VM_DISKS event:

    [root@node ~]# tail -n 1 /var/parallels/
    05-01 22:42:15.191 F /vm:28482:b3946700/ Sending question = PET_QUESTION_DELAYED_COMPACT_VM_DISKS

    if you have any other message over there - stop executing commands! Killing a running virtual machine process might lead to data corruption.

    Proceed only if you get PET_QUESTION_DELAYED_COMPACT_VM_DISKS in the last last message.

  2. Find out your Virtual Machine's UUID:

    [root@node ~]# prlctl list
    UUID                                    STATUS       IP_ADDR         NAME
    {08cd9db0-425a-649e-955c-5deec2e2c418}  starting
  3. Determine what prl_vm_app process is related to your Virtual Machine:

    [root@node ~]# ps aux | grep 08cd9db0-425a-649e-955c-5deec2e2c418 | grep -v grep
    root       29960  0.0  0.0 134858224 16232 ?     Sl   22:50   0:00 /usr/lib64/parallels-server/prl_vm_app {08cd9db0-425a-649e-955c-5deec2e2c418} {095c703b-a672-42e8-9b57-2722abd58b3b} ps
  4. Kill the process to release Virtual Machine:

    [root@node ~]# kill -9 29960
  5. Get rid of "prlctl start" processes which are starting your Virtual Machine, if any:

    [root@node ~]# ps aux | grep start | grep -v grep
    root        1941  0.0  0.0 108660  1216 ?        S    20:37   0:00 /bin/bash /etc/rc3.d/S55parallels-server start
    root       29893  0.0  0.0 474484 14624 pts/1    Sl+  22:50   0:00 /usr/lib64/parallels-server/prlctl start

    If there are any prlctl start processes related to VM in question, kill them:

    [root@node ~]# kill -9  29893
  6. Monitor VM state until it changes state from starting to stopped

    [root@node ~]# prlctl list
    UUID                                    STATUS       IP_ADDR         NAME
    {08cd9db0-425a-649e-955c-5deec2e2c418}  stopped
  7. Start Virtual Machine using Parallels Management Console

Search Words

vm starting state


stuck starting state


a26b38f94253cdfbf1028d72cf3a498b 2897d76d56d2010f4e3a28f864d69223 c62e8726973f80975db0531f1ed5c6a2 0dd5b9380c7d4884d77587f3eb0fa8ef

Email subscription for changes to this article
Save as PDF