Virtual machine in S3/S4 power mode are detected as SUSPENDED

[baby-gnu]

Description

When a VM has the power management tools and ask to suspend, it pass as pmsuspended in virsh list output.

Unfortunately, it can't be resumed by OpenNebula

``` Mon Apr 4 10:52:23 2022 [Z0][LCM][I]: Restoring VM Mon Apr 4 10:52:23 2022 [Z0][VM][I]: New state is ACTIVE Mon Apr 4 10:52:23 2022 [Z0][VM][I]: New LCM state is BOOT_SUSPENDED Mon Apr 4 10:52:23 2022 [Z0][VMM][I]: Successfully execute transfer manager driver operation: tm_context. Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: ExitCode: 0 Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Successfully execute network driver operation: pre. Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Command execution fail (exit code: 1): cat << EOT | /var/tmp/one/vmm/kvm/restore '/var/lib/one//datastores/0/802559/checkpoint' 'nebula80' '262a55de-3e36-4507-b928-7aeb7d2611c3' 802559 nebula80 Mon Apr 4 10:52:24 2022 [Z0][VMM][E]: restore: Command "set -e -o pipefail Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: # extract the xml from the checkpoint Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: virsh --connect qemu+tls://localhost/system save-image-dumpxml /var/lib/one//datastores/0/802559/checkpoint > /var/lib/one//datastores/0/802559/checkpoint.xml Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: # Eeplace all occurrences of the DS_LOCATION// with the specific Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: # DS_ID where the checkpoint is placed. This is done in case there was a Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: # system DS migration Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: sed -i "s%/var/lib/one//datastores/[0-9]\+/802559/%/var/lib/one//datastores/0/802559/%g" /var/lib/one//datastores/0/802559/checkpoint.xml Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: sed -i "s%/var/lib/one/datastores/[0-9]\+/802559/%/var/lib/one//datastores/0/802559/%g" /var/lib/one//datastores/0/802559/checkpoint.xml" failed: error: Failed to open file '/var/lib/one//datastores/0/802559/checkpoint': Aucun fichier ou dossier de ce type Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Could not recalculate paths in /var/lib/one//datastores/0/802559/checkpoint.xml Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: ExitCode: 1 Mon Apr 4 10:52:24 2022 [Z0][VMM][I]: Failed to execute virtualization driver operation: restore. Mon Apr 4 10:52:24 2022 [Z0][VMM][E]: RESTORE: ERROR: restore: Command "set -e -o pipefail # extract the xml from the checkpoint virsh --connect qemu+tls://localhost/system save-image-dumpxml /var/lib/one//datastores/0/802559/checkpoint > /var/lib/one//datastores/0/802559/checkpoint.xml # Eeplace all occurrences of the DS_LOCATION// with the specific # DS_ID where the checkpoint is placed. This is done in case there was a # system DS migration sed -i "s%/var/lib/one//datastores/[0-9]\+/802559/%/var/lib/one//datastores/0/802559/%g" /var/lib/one//datastores/0/802559/checkpoint.xml sed -i "s%/var/lib/one/datastores/[0-9]\+/802559/%/var/lib/one//datastores/0/802559/%g" /var/lib/one//datastores/0/802559/checkpoint.xml" failed: error: Failed to open file '/var/lib/one//datastores/0/802559/checkpoint': Aucun fichier ou dossier de ce type Could not recalculate paths in /var/lib/one//datastores/0/802559/checkpoint.xml ExitCode: 1 Mon Apr 4 10:52:24 2022 [Z0][VM][I]: New state is SUSPENDED Mon Apr 4 10:52:24 2022 [Z0][VM][I]: New LCM state is LCM_INIT ```

To Reproduce

1. start a virtual machine
2. execute virsh dompmsuspend --target mem one-XXXX
3. wait for OpenNebula to detect the new VM state
4. try to resume the virtual machine with onevm resume XXXX

Expected behavior

The virtual machine should be resumed with virsh dompmwakeup one-XXXX

Details

• Affected Component: Scheduler
• Hypervisor: KVM
• Version: 6.2.0-1.ce

Additional context Add any other context about the problem here.

Progress Status

• [x] Branch created
• [x] Code committed to development branch
• [x] Testing - QA
• [x] Documentation
• [x] Release notes - resolved issues, compatibility, known issues
• [x] Code committed to upstream release/hotfix branches
• [x] Documentation committed to upstream release/hotfix branches

[defekkt]

Same issue on 6.6

[brodriguez-opennebula]

OK, checked some virsh internals. Suspending the VM to S3 or S4 states depends on the default settings of the following options (opennebula doesn't set them):

  <pm>
    <suspend-to-disk enabled='yes'/>
    <suspend-to-mem enabled='yes'/>
  </pm>

If they are enabled

• a user with enough permissions inside the VM can execute a systemctl suspend or a systemctl hibernate
• virsh dompmsuspend one-$VMID mem or a virsh dompmsuspend one-$VMID disk can be executed To wake up this VMs you need to execute a virsh dompmwakeup one-$VMID In this case (power management), we get

$ virsh list
 Id   Name     State
----------------------------
 8    one-57   pmsuspended

opennebula normally suspends the VMs with a virsh suspend one-$VMID (state paused) and resumes them with virsh resume one-$VMID. That doesn't rely on the power management of the relying domain (it could work with other hypervisors). In this case we have

$ virsh list
 Id   Name     State
-----------------------
 8    one-57   paused

[onenhansen]

Suspending the VM on the host:

[root@alma8-kvm-qcow2-6-7-05dhw-1 qemu]# virsh dompmsuspend --domain one-6 --target mem
Domain 'one-6' successfully suspended
[root@alma8-kvm-qcow2-6-7-05dhw-1 qemu]# virsh list
 Id   Name    State
---------------------------
 10   one-6   pmsuspended

Checking status on frontend and resuming with updates:

[oneadmin@alma8-kvm-qcow2-6-7-05dhw-0 root]$ onevm list --no-expand
  ID USER     GROUP    NAME            STAT  CPU     MEM HOST             TIME
   6 oneadmin oneadmin test            susp    1    768M alma8-kvm-   0d 00h11
[oneadmin@alma8-kvm-qcow2-6-7-05dhw-0 root]$ onevm show 6 | grep STATE
STATE               : SUSPENDED           
LCM_STATE           : LCM_INIT            
[oneadmin@alma8-kvm-qcow2-6-7-05dhw-0 root]$ onevm resume 6
[oneadmin@alma8-kvm-qcow2-6-7-05dhw-0 root]$ onevm show 6 | grep STATE
STATE               : ACTIVE              
LCM_STATE           : RUNNING