2026.10.30a

Unraid Secretary Office 1.80.0

Added

• Mr. Backupsy shows «Retention at a glance»: a tile with the rule for all («7 d · 4 w · 6 m», Kopia's keeps, how many units differ) and a read-only table, one row per share, VM, app and the flash — local snapshots (own or inherited), Kopia, which partner office holds copies and how many, and a link to where each rule is set.

Changed

• Engine 2.82: a server whose backup settings are gone (the data folder lost or restored) no longer gets a failed run and an alert every night — the scheduled start leaves one «skipped» line that counts the nights, and the Team Lead's to-do says to set Mr. Backupsy up. His packages tile, Mr. Restori and every other reader find the backup place without settings («found without settings»). The warnings of an Apply that brings back an interrupted run's leftovers are said in the office's words.
• Mr. Backupsy's and Jack Streamer's schedules are on the same «Schedule» tile as every other desk's: the next run, the schedule in words, the last run and how it ended («Running now» while one runs); Jack's names both his jobs («Lay out: in the joint run … · Consolidate: every Sunday …»). Old bookmarks land on the new tile.
• Night watchman: a data-flow entry now says what tripped. A container or client that sends a trickle every round no longer grows one «pull» for days (the mail said «4.8 GB in 8306 min — usually at most 68 MB per hour»); the bytes a round brings are counted into the hours they really came in (a container counted anew: from its start), and the entry names the hour («3.1 GB in one hour») or the span («40 GB in 5 h 40 min») instead of the pull. «I know, thanks» makes an hour's amount normal, never the sum of a long pull. Open entries from 1.79.0 get their hour at the next round.
• The office starts clean: an agent restarted over SSH (the plugin's update, a restart by hand) no longer carries that login's session into what it starts. Before, once that login ended, a docker exec of a backup run, a restore or the Nextcloud opening could fail with «runc-process: no such file» until the next agent start.
• Mr. Restori counts a local snapshot for an app, VM or folder only when the snapshot really holds it. A share on two pools showed the other pool's snapshots too, and «newest» could point into one that had nothing of the folder. Sleeping pools are still never woken: their snapshots stay «not looked».
• Mr. Restori's page no longer says «I'm restoring» after an array stop ended the restore while the agent was away: the restore's own record decides, not the agent's last word.

Update through Unraid's plugin manager (Plugins → Check for updates). The agent restarts itself; no settings change.