Updates
LabTether can orchestrate OS package updates across your Linux fleet — plan what gets updated, preview the changes, execute across targeted nodes, and track the results. No more SSH-ing into each box to run apt upgrade.
Tip: Update orchestration currently covers the
os_packagesscope (Linux apt/dnf/pacman). Automated postchecks and rollback are not yet implemented — plan for manual verification after updates complete.
Planning updates
- Open Updates and create a new update plan.
- Select target nodes by group, type, or individual selection. Separate critical infrastructure from non-critical targets — you probably do not want to update your Proxmox host and all its VMs at the same time.
- Set the scope to
os_packages. - Run dry-run mode to preview what would change on each target. This shows you the package list without touching anything.
Before committing to a plan:
- Verify all target nodes are online and healthy.
- Check that no unresolved critical incidents are active.
- Make sure you have a recent backup or snapshot for critical systems.
make db-backupcovers the LabTether database; snapshot your VMs separately if needed.
Tip: Start with a canary approach. Update one or two non-critical nodes first, verify they survive, then expand to the rest of the fleet. The dry-run output helps you spot surprising package changes before they land.
Running updates
- Open the plan detail view.
- Start execution.
- Monitor progress as updates are applied to each target node. The plan tracks per-node status so you can see which nodes have completed, which are in progress, and whether any failed.
If a node fails during the update:
- Check agent connectivity and logs for the failed target.
- Resolve the issue manually on that node.
- Re-run the plan for remaining targets.
After updates
Automated postchecks are not yet available, so manual verification is important:
- Check telemetry and logs. Look for post-update error spikes, service restart failures, or unexpected resource changes.
- Verify nodes are reachable. All updated nodes should show connected status with a recent heartbeat.
- Spot-check key services. Open a terminal on a few updated nodes and verify that critical services came back up. Check
systemctl --failedfor any units that did not survive the update. - Compare before and after. If something looks wrong, compare pre-update and post-update logs and metrics to isolate what changed.
- Record outcomes. Add notes to the update plan or open an incident if a node needs remediation.
Warning: If a node is unstable after an update, you may need to manually revert packages or restore from a snapshot. Having a pre-update snapshot is the difference between a 5-minute rollback and a stressful afternoon.