Version 1.4.0 is available

Version 0.3.2:

  • Bug fix: --verify has no effect for some storages
  • Bug fix: prune can’t delete unreferenced chunks

Version 0.3.3:

  • Added Microsoft Azure support
  • Fixed a bug in creating directories via SFTP
  • Fixed a crash in backup when some vmdk files are not ready to be opened
  • Allow an extra colon in the sftp url so scp-style urls won’t throw an error

Version 1.0.0:

  • Added a --failure-subject option to the backup command which specifies a different email subject in presence of a backup failure
  • Added a benchmark command to measure disk and network speeds

Version 1.0.1:

  • Fixed a bug where the B2 storage can’t recover from a temporary B2 server error.

Version 1.0.2:

  • Fixed a bug that caused retention policy such as 0:m to have no effect.

Version 1.0.3:

  • Fixed a bug that caused the smtp port to be fixed at 587
  • When sending email, skip the login step if the username or password is empty

Version 1.0.4:

  • Fixed a bug in public key authentication for the SFTP storage backend

Version 1.1.0:

  • Preserve thin-provisioning when restoring disk files
  • Upload final backup file before (instead of after) removing all snapshots
  • Disable the chunk cache with a full disk
  • Retry on SFTP errors
  • Allow wildcards and excludes when specifying virtual machines to back up

Version 1.1.1:

  • Fixed a bug that caused a crash on licensee names that contain unicode characters
  • Support Wasabi storage with the url format wasabi://bucket

Version 1.1.2:

  • Fixed a crash caused by a storage password stored in the .verticalbackup/passwords file
  • Add a --parallel switch to the backup command to allow multiple backup jobs at the same time
  • Add an --exclude-disk option to exclude disks by name from backup
  • Add an environment variable VB_MAX_RETRIES to control the number of retries for the B2 storage backend
  • Fixed a bug that caused an error in opening a disk file to not display the error message
  • Fixed a bug that caused vms with spaces in the names to fail to be backed up to b2

VB_MAX_RETRIES <- Would you be so kind and allow us to set things like this in some json config instead, messing with the VMware host is bigtime pain, thanks! :slight_smile:

Version 1.1.4:

  • Fixed a bug that caused backup to a B2 bucket with spaces in the name to fail
  • Now support minio servers and generic S3 storages
  • Fixed a bug that caused prune to fail to resurrect a fossil when the corresponding chunk has been recreated
  • Fixed a bug that caused backup to abort when the licensee name contains unicode characters
  • Fixed a bug that caused check to use too much memory
  • Reopen https port when the firewall setting has been cleared due to a reboot

Version 1.1.5:

  • Fixed a bug in excluding disks by name
  • Fixed a bug that caused license update to fail
  • Release unnecessary memory to reduce memory usage for the backup command
  • Fixed a bug that prevented a non-standard sftp port from being used
  • Add a --no-chunk-listing option to the backup command to avoid listing all chunks on storage for the first backup of a VM
  • Support Amazon S3 dual stack endpoints

Version 1.2.0:

  • Optimize the in-memory chunk list to reduce memory usage
  • Fix encryption compatibility issue with Duplicacy
  • Switch to the fixed-nesting scheme adopted by Duplicacy since version 2.0.10
  • Fixed a bug in thread creation which may leave some dangling threads

Version 1.2.1:

  • Fixed a bug in determining which snapshots are new to a fossil collection (which may cause prune to mistakenly delete some fossils)
  • Moved the config file to the new format that includes a salt and the number of iterations (for compatibility with Duplicacy)
  • Fixed encryption compatibility issues with Duplicacy

Version 1.3.0:

  • Fixed out-of-memory issues when backing up vms with large disks (by reimplementing the chunk cache as a on-disk hash table)
  • Allow B2 application specific keys
  • Include .nvram files in the backup
  • Add a --snapshot-timeout option to specify longer snapshot creation operations

Version 1.4.0:

  • Implement the chunk cache as an on-disk hash table to reduce memory usage
  • Skip independent disks when backing up virtual machines
  • Use a B2 class B API call to check if a chunk already exists in the storage (instead of a class C call)
  • Avoid frequent calls to b2_authorize_account
  • Handle virtual machines with spaces in the name properly in the b2 backend
  • Gracefully exit when an upload/download error occurs

Hi @gchen
I see version 1.4.3 is current—any release notes?
Are you still updating this thread?

Version 1.4.3:

  • Add a new storage url prefix minios:// to support HTTPS connections to a minio server

The only change between 1.4.3 and 1.4.0 is the HTTPS support for minio servers. The gap in the version number was because 1.4.1 and 1.4.2 were made for testing only.