New evidence.
Once I add this (note the --json option to fwupdmgr):
===File /etc/systemd/system/fwupd-refresh.service.d/override.conf===
[Service]
DynamicUser=no
ExecStart=
ExecStart=/usr/bin/fwupdmgr --json refresh
============================================================
I get these on syslog:
fwupdmgr[3604332]: {
fwupdmgr[3604332]: "Error": {
fwupdmgr[3604332]: "Domain": "FwupdError",
fwupdmgr[3604332]: "Code": 9,
fwupdmgr[3604332]: "Message": "Metadata is up to date; use --force to
refresh again."
fwupdmgr[3604332]: }
fwupdmgr[3604332]: }
Which, in my understanding, means that fwupdmgr refuses to annoy the database
so often and returns an error because it is being called too often.
In fact pinging the database hourly as the Debian package does makes no sense
to me. So I added an override to run it weekly:
===File /etc/systemd/system/fwupd-refresh.timer.d/override.conf===
[Timer]
OnCalendar=
OnCalendar=weekly
============================================================
If I observe errors next week I will report that (hoping that I do not miss it).
-- fp
