Version 0.66

Version 0.66

• add --cbw-timeout parameter

Full Changelog: v0.65...v0.66

Version 0.65

Version 0.65

• Create fleece image with the same image options as the source image. If the fleecing image is created with the default cluster_size this could lead to backup issues if the original image uses another cluster_size option.

Full Changelog: v0.64...v0.65

Version 0.64

Version 0.64

• Add --pause option: suspend/resume VM during backup operations. Can be useful for VM's that doe not support filesystem freeze to ensure better consistency.

Full Changelog: v0.63...v0.64

Version 0.62

Version 0.62

Do not use internal qemu node names for bitmap or backup target folders
This is a breaking change for existing installations. Currently the
utility used the internal qemu node names as backup target folder. For
example, a drive specified via:

-drive node-name=disk1

was backed up to /target/disk1/[..].

This comes with several issues:

1. qemu has had multiple different ways to specify disks, there exist
   shortcuts like -hda, -drive.. or the -blockdev notations. Depending
   on the options, the configured device node names are static (if
   node-name is specified) or the node names are automatically picked
   by qemu.

   If the node-name is not specified, the automatically picked device
   node names follow a block notation, and will change between
   stop/start of the qemu process.

   In these cases, obviously the backup target folder changes and
   incremental backups may fail due to checking or existing backups in
   the wrong folders.

2. In case external snapshots are created while the virtual machine
   process is running, the active snapshot device is backed up. This
   snapshot device however, receives a new node name and the backup
   target folder would change without the user noticing. This makes it
   harder to find the backup for the right disk. Now the target folder
   stays the same.

3. The node name was also part of the bitmaps created, which can cause
   lots of other troubles if the internal assigned node names change.
   Now, the same bitmap name is used for all attached devices ruling
   out possible issues caused by vm configuration changes.

Full Changelog: v0.62...v0.63

Version 0.62

Version 0.62

• Add support for QMP TCP endpoints (-socket tcp:host:port)
• Add --progress-interval option
• Code cleanup
• Handle snapshot devices

Full Changelog: v0.58...v0.62

Version 0.58

Version 0.58

• Code cleanups
• Fix handling in case backup tool receives signal and/or backup jobs fail due to I/O errors (for example, full target filesystem):

1. Cause for failed backup job is now reported
2. If signal is received, all running jobs are cancelled and

In both cases, correctly cleanup attached fleece/snapshot/cbw devices and filters

Full Changelog: v0.57...v0.58

Version 0.57

Version 0.57

• Fixes regarding proxmox version 9, where virtual machines may use different machine types and need to be treated differently.

Version 0.56

Version 0.56

• Report reason for block job error in case backup fails

Full Changelog: v0.55...v0.56

Version 0.55

Version 0.55

• Code cleanups

Full Changelog: v0.54...v0.55

Version 0.54

Version 0.54

• Update README: add example for qcow2 data-file setting
• Code cleanups
• Print warning if VM has emulated TPM devices attached

Full Changelog: v0.53...v0.54