Version 0.18
-----------
 * Add version and parameter to log output.

Version 0.17
-----------
 * Set cleanup commands before waiting for NBD endpoint to appear,
 so qemu-nbd processes are stopped even if NBD connection might
 fail due to network constraints.

Version 0.16
-----------
 * Parse disc discard mode from libvirt disk xml and use it for the target nbd
 server.
 * Use original cluster_size of qemu images for the target images.

Version 0.15
-----------
 * Re-Define VM with new configuration on the target system after each
 synchronization so the vm configuration does always represent the verison
 from the source system.
 * Add builds for Debian related distros.

Version 0.14
-----------
 * Colorize log output, can be disabled by using NO_COLOR environment
 variable.
 * If VM is started in pause mode, skip attempt to freeze filesystems as
 qemu agent is not reachable in these cases.

Version 0.13
-----------
 * Add -start option: if VM is not running, start in PAUSED mode to allow
 replication, destroy after replication has finished.

Version 0.12
-----------
 * Use multiple concurrent processes for syncing, one worker for each disk.

Version 0.11
-----------
 * rework logging facility

Version 0.10
-----------
 * use libvirt.org/libvirtxml for parsing or modifying domain XML in case
 different name should be used.
 * move some logging statement so debug (enabled via -debug true)
 * Code cleanups
 * Thaw freezed filesystems if checkpoint creation fails.

Version 0.9
-----------
 * Use temporary images for incremental sync on target system with pointing to
 the full sync as base image. After successfull copy of the changed regions,
 commit the changes to the base image and remove the temporary qcow file.

Version 0.8
-----------
 * Get nbd maximum request size from source and target and use one that
 both backends support.
 * Enable pprof endpoint if PROFILE=development is set.
 * Start qemu-nbd process on target using --discard=unmap

Version 0.7
-----------
 * Add filesystem freeze/thaw functions using qemu guest agent, issue warning
 in case action fails.
 * Change wording for stopping the backup job

Version 0.6
-----------
 * Remove obsolete -output-dir option

Version 0.5
-----------
 * Log skipped devices

Version 0.4
-----------
 * Delete parent checkpoints if sync is successfull, keep only one checkpoint
 defined for source VM which will then be synced on next execution.

Version 0.3
-----------
 * Fix throughput shown after change to slog
 * Do not attempt any action on target system in case no dirty extents are
 found during copy.

Version 0.2
-----------
 * Check if source domain is running before attempting sync
 * Use slog for logging

Version 0.1
-----------
 * Inital Release