pi zero, 2, 3 and 4 in same cluster all recover automatically from wifi outage
but orange pi zero remains offline
can this be fixed by the balena image wifi stack ?
do you need a github bug ticket ? if so you can have full access to logs
may be related to comment
Awesome @shaunmulligan! It’s great to know someone is working on it. I don’t have any yocto knowledge but it sounds like I will be gaining some soon. Would be good to dust off the 2E. Looking forward to having it supported in the future. Regards, Michal
by cluster i mean a cluster of balena apps : one app has a zero, 2 & orange the other app has a 3 and a 4
both apps are connected to the same wifi access point and bizarrely the orange pi zero has come back on online today !
maybe the orange pi zero wifi driver only resyncs at midnight ?
It should be something else since I don’t think the orange pi zero wifi only resyncs at midnight, can you please send us all the logs you can collect so we can help investigate the issue?
14.04.20 19:34:16 (+0100) Service is already running 'main sha256:d30142851efc1dd8c622890b9abac976fd6e7ce517677e9f24b476de0192d206'
14.04.20 20:41:22 (+0100) Service is already running 'main sha256:d30142851efc1dd8c622890b9abac976fd6e7ce517677e9f24b476de0192d206'
15.04.20 02:36:43 (+0100) Service is already running 'main sha256:d30142851efc1dd8c622890b9abac976fd6e7ce517677e9f24b476de0192d206'
16.04.20 01:07:29 (+0100) Service is already running 'main sha256:d30142851efc1dd8c622890b9abac976fd6e7ce517677e9f24b476de0192d206'
root@e19fbba:~# cat /etc/hostapd/hostapd.conf
cat: /etc/hostapd/hostapd.conf: No such file or directory
root@e19fbba:~# cat /etc/dnsmasq.d/wlan0.conf
cat: /etc/dnsmasq.d/wlan0.conf: No such file or directory
root@e19fbba:~# cat /etc/network/interfaces
cat: /etc/network/interfaces: No such file or directory
root@e19fbba:~# more /var/log/syslog
more: stat of /var/log/syslog failed: No such file or directory
# cat !$
cat etc/wpa_supplicant.conf
ctrl_interface=/var/run/wpa_supplicant
ctrl_interface_group=0
update_config=1
network={
key_mgmt=NONE
}
Hi there, wpa_pairwise is a hostapd option, which is a Linux component that runs an access point as is not part of balenaOS by default. From reading this thread, you are having issues with establishing a Wi-Fi client connection to an external AP. balenaOS uses NM to manger network connections and we have a lot of examples how to customise its configuration here.
The message you are seeing in the logs has probably more specific to the kernel version/wifi module and possibly to do with a wider hardware issues, such as elevated device temperature and high CPU utilisation.
Have you checked dmesg output for other hints as well as run the device diagnostics from the dashboard?
We would need to see the logs from the device at the time the reconnect happens to help troubleshoot further. Are you able to provide these please? /var/log/messages on the host is likely where we will see something…
No worries, just in case there is nothing evident in the logs, another suggestion would be to try and disable the on-board WLAN on the OPi, since there are reports of it being unstable and instead use an external USB adapter. Failing that, another workaround may be to wire up a watchdog type script to run periodically on the device, checking for network connectivity and manually reconnecting if detected offline.
all good ideas but we’re talking about the IOT here and maybe i want to deploy a few 1000 of these into the wild i can just move on to another hw board vendor rather than mess around with having to buy EXTRA hw or rolling my own (unsupported) sw workarounds
so ideally it would be great if the balena drivers auto reconnected wifi for me and maybe we will discover a flaw in the drivers that can improve wifi for all balena targetted devices
FYI latest update its 7:25 and its still offline my midnight theory could be right
TBH if it went offline and stayed offline i would probably chuck it in the bin and buy banana pis instead its just strange that it CAN come back online