[lunar] [rpi] smartmontools.service fails to run

[Paul Larson]

I don't think this service should normally be active for most raspberry PIs, but it is, and fails. This gets picked up by some automated tests that we normally run for certification testing as an error. Here's the output from journalctl -u smartmontools:
Mar 30 14:22:09 ubuntu smartd[640]: smartd 7.3 2022-02-28 r5338 [aarch64-linux-6.2.0-1003-raspi] (local build)
Mar 30 14:22:09 ubuntu smartd[640]: Copyright (C) 2002-22, Bruce Allen, Christian Franke, www.smartmontools.org
Mar 30 14:22:09 ubuntu smartd[640]: Opened configuration file /etc/smartd.conf
Mar 30 14:22:09 ubuntu smartd[640]: Drive: DEVICESCAN, implied '-a' Directive on line 21 of file /etc/smartd.conf
Mar 30 14:22:09 ubuntu smartd[640]: Configuration file /etc/smartd.conf was parsed, found DEVICESCAN, scanning devices
Mar 30 14:22:09 ubuntu smartd[640]: In the system's table of devices NO devices found to scan
Mar 30 14:22:09 ubuntu smartd[640]: Unable to monitor any SMART enabled devices. Try debug (-d) option. Exiting...
Mar 30 14:22:09 ubuntu systemd[1]: smartmontools.service: Main process exited, code=exited, status=17/n/a
Mar 30 14:22:09 ubuntu systemd[1]: smartmontools.service: Failed with result 'exit-code'.
Mar 30 14:22:09 ubuntu systemd[1]: Failed to start smartmontools.service - Self Monitoring and Reporting Technology (SMART) Daemon.

ProblemType: Bug
DistroRelease: Ubuntu 23.04
Package: smartmontools 7.3-1
ProcVersionSignature: User Name 6.2.0-1003.3-raspi 6.2.2
Uname: Linux 6.2.0-1003-raspi aarch64
ApportVersion: 2.26.0-0ubuntu2
Architecture: arm64
CasperMD5CheckResult: unknown
CloudArchitecture: aarch64
CloudBuildName: server
CloudID: nocloud
CloudName: unknown
CloudPlatform: nocloud
CloudSerial:

CloudSubPlatform: seed-dir (/var/lib/cloud/seed/nocloud-net)
Date: Thu Mar 30 20:19:41 2023
ImageMediaBuild: 20230329
SourcePackage: smartmontools
UpgradeStatus: No upgrade log present (probably fresh install)

[Ubuntu QA Website (ubuntuqa)]

This bug has been reported on the Ubuntu ISO testing tracker.

A list of all reports related to this bug can be found here:
https://iso.qa.ubuntu.com/qatracker/reports/bugs/2013393

[Dave Jones (waveform)]

This is odd; I can't replicate this with the current lunar preinstalled server for rpi images, either the arm64 or armhf ones (admittedly I've only tried the 2GB Pi 4 so far). "pgrep smartd" returns nothing, "journalctl -u smartmontools" reports "No entries". Do you have something like a customized cloud-init configuration which is installing it?

[Christian Franke (christian-franke)]

Starting with smartmontools 7.3, you could add '-q nodev0' to smartd command line.
Then smartd will exit with status 0 instead of 17 (default '-q nodev') if there are no devices to monitor. Systemd should no longer report this as a failed service.

[Dave Jones (waveform)]

Double-checked the manifest (http://cdimage.ubuntu.com/ubuntu-server/daily-preinstalled/current/lunar-preinstalled-server-arm64+raspi.manifest) and smartmontools isn't in it. Something else (which I can only guess is cloud-init at this stage) must be pulling it in for pwlars.

[Paul Larson (pwlars)]

Ok, it looks like it may have come from the test suite we used. So it's still a different behavior from what we saw in the past, as I don't think this is a new dependency. Perhaps something may have changed in the default config that triggered this? But for purposes of iso testing, indeed it doesn't seem relevant.