Open antortjim opened 3 years ago
We do see wifi issues from time to time. For some mysterious reason, they tend to be concentrated in time or space. For the past several months I believe things have been more stable. I never managed to crack them down but one thing that has been in my list of things to do is to switch the platform away from DHCP and use static IPs instead. This would have the advantage to quickly map the ethoscope to the IP address.
I will give that a try asap, and get back here when I know something new. Thank you @ggilestro
Dear ethoscope community.
I have encountered today an issue that I already encountered in the past, but now I have been able to gather more information about it. I know it is not an ethoscope specific problem, but more of a RPi thing, but still very useful for the community to be aware of.
The problem is a particular ethoscope keeps connecting and disconnecting from the network. It keeps reconnecting to the network during an experiment. But I think it doesn't have the issue during a following experiment (I still need to confirm this).
I have checked the logs of netctl-auto by calling
journalctl -ru netctl-auto@wlan0
(see logs at the end)Basically the same block of logs from
DUID ...
todhcpcd exited
keep repeating every few minutes. On the block it is very clear that a particular program is getting sent the SIGTERM signal, which kills the program. These kills match the times when the ethoscope goes offline. I checked the PID and the process being killed iswhich indeed makes sense. That is the program in charge of setting the WiFi connection.
My question is WHAT is causing this process to be killed or at least WHY does it happen. I have tried disabling the power saving mode with
but it keeps happening
I have also checked the logs of
systemd-networkd
, which contain theGained carrier
andLost carrier
messages at time matching the logs ofnetctl-auto
. So it is definitely connected.I don't know if this is caused by a defective RPi or maybe an overloaded network. But I can say when it happens, it keep happening on the same ethoscope. I would expect this problem to happen to a different ethoscope every time if the network was overloaded. I wonder whether an ethernet connection (via cable) would patch this problem....
systemd-networkd
netctl-auto@wlan0
Thank you, Antonio