Backups

restic every night, Proxmox dumps every week, and the restores that prove they work.

TIL

lsblk shows the serial of the disk you are about to pull

lsblk takes a list of columns, and SERIAL is one of them. The serial is printed on the disk’s label, so it tells you which of four identical disks is sdc.

$ lsblk -d -o NAME,SIZE,SERIAL
NAME  SIZE SERIAL
sda   3.6T WX12D7A0K4L2
sdb   3.6T WX12D7A0K9F1
sdc   3.6T WX12D7A0J3N8
sdd   3.6T WX12D7A0K1C5

sdc reported 8 reallocated sectors this week. The new disk arrives on Monday, and I will pull the one labelled WX12D7A0J3N8, not the one in the third slot.

Incident

Backups stopped for 31 hours

The NAS ran out of space on Thursday night and restic failed at 04:00. The alert went to an address I stopped reading in spring.

Resolved Major

A terminal on the NAS: the three newest restic snapshots of pve-1, from 2026-11-03, 2026-11-04 and 2026-11-05 at 04:00, and df showing /srv at 7.3T of 7.3T, 100 % used.

TIL

systemd-analyze calendar shows when a timer fires next

systemd-analyze calendar reads an OnCalendar= expression and prints the next time it matches, in local time and in UTC.

$ systemd-analyze calendar '*-*-* 04:00:00'
  Original form: *-*-* 04:00:00
Normalized form: *-*-* 04:00:00
    Next elapse: Fri 2026-08-21 04:00:00 CEST
       (in UTC): Fri 2026-08-21 02:00:00 UTC
       From now: 9h left

I check every new expression with it before a timer gets it. --iterations=3 lists the next three matches, which is how I found that the Sunday prune was set for Saturday.

Buy Logbuch