UPS monitoring, Nextcloud cron, container image bumps
ups role: NUT server on home.debyl.io for the CyberPower PR1500RT2U that backs both it and truenas.localdomain, with staged shutdown (TrueNAS sheds at t+2min, host at 10% charge) and best-effort IPMI power-on when mains returns. The 10% threshold leans on ignorelb + override.battery.charge.low rather than a custom poller, because CyberPower asserts its own low-battery flag far too early. Credentials come from vault vars; nothing sensitive is templated in the clear. Nextcloud background jobs: both instances have backgroundjobs_mode "cron", which expects an external caller every ~5 minutes, and nothing was calling. The personal instance had not run a background job since 2026-05-14 and skudak since 2024-11-20. Consequently trash and file versions never expired, stale chunked uploads accumulated, calendar reminders never fired, and nextcloud.log was never rotated -- which quietly made the existing log_rotate_size cap inert. Added a systemd timer per instance, skipping cleanly when the container is down or in maintenance so deploy windows don't show up as failed units. Trash retention on the personal instance: the default "auto" only expires when disk space demands it, so 66 GB of >30-day deletions sat on a host with 1.3 TB free -- effectively unbounded. "auto, 30" makes the 30-day expiry unconditional while still purging early under pressure. Image bumps: nextcloud 33.0.0 -> 34.0.2 (both cloud and skudak-cloud) greg-time-bot 3.9.25 -> 3.10.0 fulfillr 20260723.2044 -> 20260728.2155 (prod and dev) The fulfillr bump records what is already deployed: both containers were rolled to that image on 2026-07-29 for SCRUM-156 (digital product releases + customer update campaign). Committing it keeps the repo from claiming an older tag than the host is actually running, which would otherwise roll fulfillr backwards on the next clean-checkout deploy. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
This commit is contained in:
Bastian de Byl
@@ -0,0 +1,114 @@
|
||||
# ups
|
||||
|
||||
UPS monitoring and staged shutdown for the home rack.
|
||||
|
||||
A CyberPower PR1500RT2U (`0764:0601`) is cabled by USB to `home.debyl.io` and
|
||||
backs both that host and `truenas.localdomain` (Dell PowerEdge R415). This role
|
||||
makes `home.debyl.io` the NUT server and gives it the ability to power TrueNAS
|
||||
back on over IPMI.
|
||||
|
||||
## Outage sequence
|
||||
|
||||
| When | What happens | Driven by |
|
||||
| --- | --- | --- |
|
||||
| t+0 | UPS goes on battery, `ONBATT` logged to journald → Graylog | `upsmon` |
|
||||
| t+2min | TrueNAS shuts itself down cleanly, shedding ~200 W | TrueNAS UPS service, Slave mode, `Shutdown Timer 120` |
|
||||
| 10% charge | `home.debyl.io` shuts itself down and tells the UPS to cut its output | `upsmon` `SHUTDOWNCMD` + `/lib/systemd/system-shutdown/nutshutdown` |
|
||||
| mains returns | UPS re-energizes; both machines power themselves back up | R415 `always-on` restore policy; `home.debyl.io` BIOS *After Power Loss → Power On* |
|
||||
| mains back +5min | Best-effort IPMI power-on, if a dedicated iDRAC is ever fitted | `upssched` → `ups-restore.sh` |
|
||||
| host boot | Same restore check, for the deep-drain case | `ups-restore.service` |
|
||||
|
||||
The 10% threshold is not a custom poller. CyberPower asserts its own low-battery
|
||||
flag around 20–35%, so `ups.conf` sets `ignorelb` plus
|
||||
`override.battery.charge.low`, and stock `upsmon` fires at exactly the
|
||||
configured percentage.
|
||||
|
||||
Likewise, the 2-minute TrueNAS shed is TrueNAS's own native "shutdown timer"
|
||||
setting — no SSH key and no shutdown script from this side.
|
||||
|
||||
## Powering TrueNAS back on
|
||||
|
||||
Neither Wake-on-LAN nor IPMI works on this box today, so restore is done with
|
||||
the chassis power restore policy instead.
|
||||
|
||||
**Wake-on-LAN is out.** `bce0` advertises no WOL capability (`ifconfig -m bce0`
|
||||
has no `WOL_MAGIC`) and both NICs are bonded into an LACP `lagg0`.
|
||||
|
||||
**IPMI is out too, for now.** The iDRAC is in `shared / LOM1` mode and the
|
||||
Enterprise card that would provide a dedicated management port is not fitted:
|
||||
|
||||
```
|
||||
$ ipmitool sdr elist | grep -i idrac
|
||||
iDRAC6 Ent Pres | 70h | ok | 7.1 | Absent
|
||||
```
|
||||
|
||||
A shared-LOM iDRAC6 Express has no standby power. Measured directly: with the
|
||||
chassis powered off the BMC does not even answer ARP, and it only reappears
|
||||
~200s into POST, at the moment the host brings the NIC link up. That is a
|
||||
hardware limitation, not a switch or BIOS problem.
|
||||
|
||||
**So restore works like this instead.** `ipmitool chassis policy always-on` is
|
||||
set on the R415. On a deep outage `home.debyl.io` halts at 10% and NUT's
|
||||
shutdown hook tells the UPS to cut its output; when mains returns the UPS
|
||||
re-energizes, the R415 sees AC and boots itself.
|
||||
|
||||
The gap is the medium outage — mains returns after TrueNAS has shed but before
|
||||
the battery reaches 10%. The UPS never cuts power, so TrueNAS stays off and
|
||||
needs a manual power button press. Fitting a used iDRAC6 Enterprise card and
|
||||
running `ipmitool delloem lan set dedicated` (plus a cable to the dedicated
|
||||
port) closes that gap, and `ups-restore.sh` starts working with no code
|
||||
changes — it is already deployed and simply logs and exits while the BMC is
|
||||
unreachable.
|
||||
|
||||
## One-time setup outside Ansible
|
||||
|
||||
These are not managed by this role.
|
||||
|
||||
**iDRAC (already done, via `ipmitool` on TrueNAS):**
|
||||
|
||||
```
|
||||
ipmitool lan set 1 ipsrc static
|
||||
ipmitool lan set 1 ipaddr 192.168.1.12
|
||||
ipmitool lan set 1 netmask 255.255.255.0
|
||||
ipmitool lan set 1 defgw ipaddr 192.168.1.1
|
||||
ipmitool lan set 1 access on # was disabled - nothing answers without this
|
||||
ipmitool channel setaccess 1 2 callin=on ipmi=on link=on privilege=4
|
||||
ipmitool user set password 2 '<idrac_password>'
|
||||
ipmitool chassis policy always-on # this is what restores power after an outage
|
||||
```
|
||||
|
||||
`idrac_password` is also the iDRAC web UI password for `root` - they share a
|
||||
user database.
|
||||
|
||||
**TrueNAS UI → Services → UPS** (enable + start automatically):
|
||||
|
||||
| Field | Value |
|
||||
| --- | --- |
|
||||
| UPS Mode | Slave |
|
||||
| Remote Host | `192.168.1.10` |
|
||||
| Remote Port | `3493` |
|
||||
| Identifier | `cyberpower` |
|
||||
| Monitor User | `truenas` |
|
||||
| Monitor Password | vault `nut_truenas_password` |
|
||||
| Shutdown Mode | UPS goes on battery |
|
||||
| Shutdown Timer | `120` |
|
||||
| Shutdown Command | `/sbin/shutdown -p now` |
|
||||
| Power Off UPS | unchecked |
|
||||
|
||||
**BIOS on `home.debyl.io`** (Lenovo 10MR0004US): Power → *After Power Loss* →
|
||||
**Power On**. On a full drain the NUT shutdown hook
|
||||
(`/lib/systemd/system-shutdown/nutshutdown`) tells the UPS to cut its own
|
||||
output; this BIOS setting is what brings the host back when mains returns.
|
||||
|
||||
## Vault keys
|
||||
|
||||
`nut_upsmon_password`, `nut_truenas_password`, `idrac_password`
|
||||
|
||||
## Operating
|
||||
|
||||
```
|
||||
upsc cyberpower # full UPS status
|
||||
upsc cyberpower battery.charge
|
||||
sudo /usr/local/bin/truenas-power.sh status|on|soft|off
|
||||
journalctl -t ups-restore -t ups-sched
|
||||
```
|
||||
Reference in New Issue
Block a user