Everything started out well. Then I switched to my main pc. After an hour I realized that the coordinator is no longer capturing messages.
After some googling I came to a realization that it is 99% related to usb autosuspend. So I added usbcore.autosuspend=-1 to grubâs config. And the problem seemed to go away (on the first day I reproduced it a few times to be sure that it wasnât a onetime glitch).
And, just now, after 2 days of silence,I caught it again(all of a sudden). Zigbee2mqtt, which has a front end, which I accessed from my main system, was working well. But did not capture any messages (like the first times).
So I switched my monitor input to the lab pc, and decided to look whats going on. At that moment everything started working again (seems to be the moment I logged in or something).
I have done another round of searching what could be done, and, for the moment, I disable autosuspend by timeout from the settings.
Any idea how to address this further?
A noobish question, but is there a possibility to somehow âpingâ a device/usb port in order to always keep it online (and eliminate the scenario of the device getting suspended), which I could wrap in a script and set a cronjob for it?
With the release of fedora (approx 38 if I remember correctly) fedora switched to automatically suspend the desktops after 15 minutes idle. This caused problems for many users who used their systems as a 24hr online server. Idle is defined as no keyboard or mouse activity and does not include remote connections.
Suspend/sleep/hibernation can be disabled with a couple of changes. see the following links for details.
and
On my system (f43) I do not have the /etc/systemd/sleep.conf.d directory but instead placed the sleep.conf file as shown below. I believe that you could create the /etc/systemd/sleep.conf.d/ directory and place your sleep.conf file there as well.
You also could mask the appropriate services and have the same effect system wide and not limited to your user environment.
Those services are systemd-suspend, systemd-sleep, systemd-suspend-then-hibernate, & systemd-hibernate.
They can be made inactive with sudo systemctl mask systemd-{suspend,sleep,suspend-then-hibernate,hibernate}.service.
Well, I am now testing the âmaybe I needed to rebootâ after those changes. I have made the changes more than 24 hours ago. All seemed fine. An hour ago I started to notice delayed reaction to button presses being received by the receiver.
And, once again, the moment I turned the monitor on, moved the mouse, all started working like a clock.
Well, I am now testing the âmaybe I needed to rebootâ after those changes. I have made the changes more than 24 hours ago. All seemed fine. An hour ago I started to notice delayed reaction to button presses being received by the receiver.
And, once again, the moment I turned the monitor on, moved the mouse, all started working like a clock.
Just to confirm, the changes, that were to be made, are:
Disable suspend in KDEâs Power Management.
/etc/systemd/sleep.conf
Correct? (I also found a âblock suspendâ in the tray, and created a
I am also exercising one interesting moment - this thing has happened 2-3 times already. But never in the morning - always in the evening.
I have 3 pcâs connected to my monitor - my home pc(dp), my working pc(hdmi), and the lab one(also hdmi). And evening is the moment when I suspend my working pc (Ubuntu), switch to my main(win10). And, later, I turn it off and move to the tv. And this is the time period, where I usually find that the light controls are unresponsive (controller fails to capture messages).
Although today it was strange, and I first noticed that the few clicks were processed with a delay. After it was zero responsivness. But all returned to normal once I switched the monitor input to the lab pc (and the kb and mouse to bt⌠and made a mouse movement).
Have a hunch there is some suspend/sleep trigger happening by the monitor going to sleep related event(Iâve seen google bring up this topic a few times).