7.10.6.9 — August 2026

Services

• Added alerting across the virtualisation and backup platforms, so a failed backup, a failed integrity check or a hardware fault now pushes to a phone instead of an unread local mailbox.
• Added processor and memory to every workload, reclaimed from the management appliances the old platform required and the new one does not.
• Added a verified network failover test: live links were physically pulled and every service stayed reachable, instead of the redundancy being assumed to work.
• Added a weekly all-clear from every machine in the platform, so a quiet week now confirms the backups ran instead of leaving it an open question.
• Added the guest tooling the backup service was missing, so it can be snapshotted with its filesystems quiesced and shut down cleanly on power loss.

Stability

• Fixed automatic start-up, switched off for the hardware move and never restored, which meant the machines would have shut down cleanly on a power cut and then never come back.
• Fixed the last failing battery backup unit, so all three machines now shut themselves down safely on utility power loss rather than two out of three.
• Retrofitted the backup service onto the same automated configuration as the rest of the estate, so it can be rebuilt from a clean install rather than from memory.
• Retrofitted the cluster database onto power-loss-protected storage, so an abrupt power cut cannot corrupt the record of what the cluster is meant to be running.
• Fixed the backup service not starting by itself, which left it off after a host restart until somebody noticed.

Tracks in this release: Services · Stability