| Name | Modified | Size | Downloads / Week |
|---|---|---|---|
| Parent folder | |||
| SOURCE-CODE.txt | 2026-10-28 | 1.7 kB | |
| build-winpe.ps1 | 2026-10-28 | 3.8 kB | |
| install.sh | 2026-10-28 | 8.6 kB | |
| manifest.json | 2026-10-28 | 5.2 kB | |
| manifest.json.sig | 2026-10-28 | 89 Bytes | |
| backupzit-server_0.44.2_amd64.deb | 2026-10-28 | 64.1 MB | |
| backupzit-server-0.44.2-1.x86_64.rpm | 2026-10-28 | 64.5 MB | |
| backupzit-repo_0.44.2_amd64.deb | 2026-10-28 | 3.3 MB | |
| backupzit-agent-0.44.2-1.x86_64.rpm | 2026-10-28 | 9.5 MB | |
| backupzit-agent-0.44.2-x64-legacy.msi | 2026-10-28 | 10.1 MB | |
| backupzit-agent-0.44.2-x64.msi | 2026-10-28 | 10.0 MB | |
| backupzit-agent_0.44.2_amd64.deb | 2026-10-28 | 9.4 MB | |
| backupzit-repo-0.44.2-1.x86_64.rpm | 2026-10-28 | 3.3 MB | |
| BackupZit 0.44.2 source code.tar.gz | 2026-10-28 | 5.9 MB | |
| BackupZit 0.44.2 source code.zip | 2026-10-28 | 6.5 MB | |
| README.md | 2026-10-28 | 3.1 kB | |
| Totals: 16 Items | 186.6 MB | 0 | |
Proxmox VE: incremental backups read only what changed
Running VMs on LVM-thin, ZFS, Ceph RBD or qcow2 now get changed block tracking (QEMU dirty bitmaps). From the second backup on, only the areas written since the previous backup are read: in our test 92 MiB instead of 9 GiB, 6 seconds instead of 44. After a VM was stopped, restarted or moved to another node, the next backup reads its disks completely once. The run page shows which disks were read as changed blocks only. Snapshots with RAM, hibernation and live migration keep working with the tracking in place.
Fixes for Proxmox VE 9
- Backups of running VMs with qcow2 disks failed on Proxmox VE 9 (its qemu-nbd no longer has
--force-share). - The guest state is asked from
qm/pctdirectly: right after a migration or a resume the cluster status could still say "stopped", and the backup then ran without freezing the file systems. - Restoring or instantly recovering a VM to the next free ID right after a guest with that ID was removed failed with "guest exists".
Back up the Proxmox host configuration with one click
On Backup jobs, choose a Proxmox node's agent and Files and folders: Fill in the host configuration enters /etc, /var/lib/pve-cluster and /root. The agent makes a consistent copy of the configuration database config.db before each backup; sqlite3 is no longer needed. Restore steps are in the Proxmox guide.
Find backups on local and USB disks
Find backups now also works for local and USB disk targets: choose the agent of the computer the disk is connected to, it searches the disk and the backups can be added like those on network storage.
Clock warnings
Agents send their time with every poll. When a computer's clock differs from the console's by 2 minutes or more, the dashboard and the Agents page say so (S3 and Azure refuse requests from 15 minutes on).
Progress
Restores show a progress bar too (with the percentage when a whole backup is restored). The estimate for file backups uses both the files and the size of the previous backup, so it no longer disappears when a backup has more files than the last one.
Deleting backups asks for the password or the two-factor code
Removing an agent together with its backups (or deleting the backups of a removed agent) now needs, besides the name of the computer, your password — or the current code from your authenticator app when two-factor sign-in is on — so a stolen browser session alone cannot delete backups. A second user who approves such a removal confirms the same way. Users who sign in with SSO need two-factor sign-in for it.
Hardening
- Find backups on a local or USB disk accepts the search result only for the searched storage location.
- Cross-Origin-Embedder-Policy header on all pages.
- The appliance no longer answers ICMP timestamp requests.
The user guide is updated in all 14 languages. Update the console and then the agents (Agents → Update all).
Releases are signed with the BackupZit release key; consoles verify them before installing.