VM created from a backup keeps the backup's data volume device ids on a config drive network and cannot start
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Newbie friendliness
- 48/100
- Issue type
- Bug
- Clarity
- Mostly clear
- Activity status
- Active
- Tech stack
- java
- Domain
- backend, infrastructure
Research direction
Start by tracing createVMFromBackup and how datadisksdetails carries backup volume device IDs into VM creation. Compare the config-drive network path and the documented q35 SCSI device layout, including the vm.iso.max.count=2 case; the payload names no source files or tests. Done when backup-created VMs on config-drive networks avoid reserved devices and start successfully across the reported cases.
Written by the indexing model from the issue text.
Description
Problem
createVMFromBackup takes the device ids of the data volumes from the backup. A backup of a VM without config drive can have a data volume on a device id that is reserved on a config drive VM (4, or 5 on main). When the VM is created from that backup on a config drive network, the volume keeps that device id, and on q35 (SATA CD-ROMs, SCSI data disks) its target is the same as a CD-ROM. The VM ends in Error state with XML error: target 'sde' duplicated (4.22) or target 'sdf' duplicated (main, only when vm.iso.max.count=2). The same backup restored on a network without config drive gives a running VM.
Not caused by the device id fixes): it also fails on 4.22.1.1 and on 4.23.0.0. On 4.23.0.0, a VM created from a backup with data volumes on 1, 2, 5, 6 on a config drive network also cannot start with vm.iso.max.count=2 (target 'sde' duplicated between the empty CD-ROM and the config drive ISO), and with the 4.23 device ids (1, 2, 4, 5) the source VM itself cannot start.
Versions
Fails before the device id fixes:
- 4.22.1.1: deviceid 4 kept, the VM cannot start (default vm.iso.max.count=1).
- 4.23.0.0: data volumes on 1, 2, 5, 6: the VM starts with vm.iso.max.count=1 and cannot start with vm.iso.max.count=2.
Still fails after the device id fixes (this path is not covered by them):
- 4.22 build with the device id fix (4.22.2.0-SNAPSHOT): same as 4.22.1.1.
- main (24.0.0-SNAPSHOT): same as 4.23.0.0 (starts with vm.iso.max.count=1, cannot start with 2).
Environment:
KVM on OL 9, qemu-kvm 7.2.0, libvirt 9.0.0, machine type pc-q35-rhel9.2.0 (template settings rootDiskController=scsi, kvm.guest.os.machine.type=pc-q35-rhel9.2.0, UEFI=LEGACY). NFS primary storage, NAS backup provider with an NFS repository.
The steps to reproduce the bug
- Create two L2 networks: N1 with DefaultL2NetworkOffering, N2 with DefaultL2NetworkOfferingConfigDrive.
- Deploy a q35 scsi VM on N1 and attach data volumes: on 4.22, 3 volumes (device ids 1, 2, 4); on main, 4 volumes (device ids 1, 2, 5, 6).
- Assign the VM to a NAS backup offering and take a backup. The backup volume list has the same device ids.
- (main only) Set vm.iso.max.count to 2.
- Create a VM from the backup on N2 (createVMFromBackup with networkids=N2).
- The job fails with 530 "Unable to orchestrate the start of VM instance". The VM is in Error, the data volume is on deviceid 4 (or 5), and the management server log has "Unable to start VM on Host ... XML error: target 'sde' (or 'sdf') duplicated" for each host.
- Control: create a VM from the same backup on N1. It is Running.
Workaround
Tested on 4.22: createVMFromBackup with datadisksdetails and device ids 1, 2, 5 gives a running VM.
What to do about it?
Expected: on a config drive network, the data volumes of a VM created from a backup do not use the device ids reserved for the CD-ROMs and the config drive ISO, and the VM starts.
- Dominant language
- Java
- Stars
- 3.1k
- Forks
- 1.4k
- Avg merge
- 7d 4h
- Merged PRs (30d)
- 13
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from apache/cloudstack
-
bug
Difficulty 2/5 1-3 hours Newbie friendliness 76/100
apache/cloudstack#14719 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 77/100
apache/cloudstack#14715 ·
Maintainers usually reply within 1 day
-
resizeVolume of a running system VM's volume is accepted but the VM never sees the new sizePossibly taken @bhouse-nexthop claimed this 7 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
apache/cloudstack#14645 · 1 comment ·
Maintainers usually reply within 1 day
-
VR reports a load balancer change as applied when haproxy.cfg cannot be replacedPossibly taken @bhouse-nexthop claimed this 7 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
apache/cloudstack#14633 · 1 comment ·
Maintainers usually reply within 1 day
-
restartNetwork reports a ResourceUnavailableException as an unexpected error (530), unlike restartVPC (534)Possibly taken @bhouse-nexthop claimed this 7 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
apache/cloudstack#14631 · 1 comment ·
Maintainers usually reply within 1 day
All issues in apache/cloudstack
Similar issues
-
Difficulty 1/5 Under an hour Newbie friendliness 75/100
SAP/ai-sdk-java#1206 ·
Maintainers usually reply within 1 day
-
enhancement good first issue insight-module
Difficulty 2/5 1-3 hours Newbie friendliness 86/100
sdlc-pro/spring-lens#606 ·
Maintainers usually reply within 1 day
-
quality-assurance
Difficulty 2/5 1-3 hours Newbie friendliness 62/100
Paul-Austin-Oswego-CSC480-HCI521/gift-app#252 ·
Maintainers usually reply within 1 day
-
backlog kind/bug
Difficulty 2/5 1-3 hours Newbie friendliness 83/100
TencentBlueKing/bk-job#4627 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 63/100
openrocket/openrocket#3460 ·
Maintainers usually reply within 1 day