Skip to content

v1.1.2 — container backup and clone from snapshot

Latest
Latest

Choose a tag to compare

View all tags
@cagamps cagamps released this 22 Oct 17:40
· 1 commit to master since this release
v1.1.2
08d9cc5
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Highlights

Container backups now work in snapshot mode, the default for scheduled backup jobs. Before this release, backing up a running container on QuantaStor storage failed with snapshot activation not supported on quantastor storage. VM backups were not affected.

Fixed

  • Container backup in snapshot mode (vzdump --mode snapshot, and backup jobs to PBS or any other target). The plugin now activates, maps and deactivates a QuantaStor snapshot as its own volume, so vzdump can mount the temporary snapshot read-only. It also logs out of the snapshot before deleting it.
  • pct clone --full --snapname, cloning a container from a snapshot, failed with the same error and now works.
  • Disk creation could fail with A storage volume with the specified name ... already exists [err=9]. A volume created moments earlier can take a couple of seconds to appear in QuantaStor's volume list, so a new disk could be given a name that was already taken. When the plugin chooses the disk name itself, it now moves on to the next free number. A volume list that can't be read is reported as an error instead of being treated as empty.

Added

  • qm clone --full --snapname, a full clone of a VM from a snapshot. The clone contains the snapshot's data. Linked clones still require a template.

Known issues

  • First start of a restored VM with a cloud-init drive. A VM restored from backup has no cloud-init volume yet, and its first start fails with qemu-img: Could not open ''. That failed start creates the volume, so starting the VM again works.
  • An occasional container backup failure in snapshot mode. On current QuantaStor releases, a backup run can occasionally fail with iSCSI login failed … exit code 21 when QuantaStor is still setting up the temporary snapshot. The next run succeeds, and no data is affected. A QuantaStor fix is in progress.
  • Reassigning a disk to another guest ("Reassign Owner", qm disk move --target-vmid, pct move-volume --target-vmid) is not supported and is refused.

Install / upgrade

apt install ./pve-storage-quantastor_1.1.2-1_all.deb

Upgrading from any 1.x release uses the same command. Existing storages and running guests are unaffected.

Validated

The full compatibility matrix on this package, against a QuantaStor 6.9 HA pair with two networks and a 2-node PVE 9.2 cluster, single-path and multipath:

  • Snapshot-mode backup of a running container with two volumes, then restore and start.
  • Container clone from a snapshot.
  • VM clone from a snapshot. The clone contains the snapshot's data, not the disk's later contents.
  • VM snapshot-mode backup, then restore and start.
  • After each case, no snapshot sessions, multipath maps or volumes were left.
  • The VM and container lifecycle, two-node migration, a QuantaStor service restart under I/O, and HA failover and owner reboot with 0 guest I/O errors, single-path and multipath.
  • Upgrades to 1.1.2 from every earlier release (v0.2.0 through v1.1.1), including in place with guests running.

SHA256

c8fdb91f2e51abc58d9e13133c2b5f9e70cd8d547db7967cd395e2f9ee0ede5c

Assets 3
Loading