jc77 wrote: > > Any suggested changes to script to make keep this one from going > completely offline every 12 hours or so? [Back1] ... > This Radio needs to be restarted after 2-3 days [Kitchen] ... > These two stay online for about a week before needing to be restarted: > [Floater & Back2] ... > Wow, this is a lot. Thanks for the report. This is the first report of a radio requiring a reboot to reconnect since a quick reset related bug caused the script to drop out last month, and you have 4 of them!. There is something happening that is out of the ordinary in your environment. Here, the only outages have been from hardware (power supply connector) failures.
To troubleshoot this, use the ::-d /etc/log/:: to save the log files where they will not be lost after a reboot, and examine the files as soon as you notice the radio off. Generally, the script will keep trying to reconnect forever, but it limits each trial to a specified amount of time before trying again. If it tries again too soon, a possible recovery is aborted, a new cycle begins, and the radio does not recover. This was a problem with weak signals and heavy interference in earlier versions, where the retry time was fixed at a relatively short time. Later versions like you are using have an adaptive retry time that has an upper limit. You might be hitting that upper limit. On the other hand, something else may be happening (e.g., the script may be exiting), so looking at the logs after a reboot is important. You can zip them and upload them into a new GitHub issue. None of your radios have any Options showing. The 'slow' web server is recommended for monitoring. Change launch in /etc/init.d/rcS.local to: :: /etc/wlanpoke/wlanpoke.sh -W slow -d /etc/log/ &:: and check out http://Back1:8080 (use Back1's IP address if needed). Also, ncat is handy. If you are not using it, disable tcp logging with the ::-x:: option. Of your radios, only "Floater" is showing extended full reset recovery times (possibly from going out of range of the access point) and it stays on for a week. All the others recover right away. You can increase this time by editing the script (no command line option yet) variables FRWaitSecsMax and FRWaitStepPct, increasing one or both. Your 'Back1' radio reports a signal level of -70 dBm. It may be just out of range of the AP given your environment. "Floater" had a very high signal level at the time of measurement, but perhaps that unit had also been in the hinterlands for a while, which would explain the repeated failure of the full resets to work, as Floater recovered when brought into a strong signal area. (Does Back1 recover without rebooting when moved into a high signal area?) Back1 is failing the most often, so it makes sense to concentrate on this unit first, and perhaps Floater as well, as it exhibits the interesting increasing recovery time steps, which might also being experienced by Back1, but the logs are lost. It would also be useful for you to track the radios as a whole to see about time of day phenomena. The excellent free Wireless Network Watcher and NetworkConnectLog from Nir Software on your Windows pc will capture the status of the devices on your LAN. This would give you an important "last seen" time to correlate with your log files. And you might try the latest software in the development branch to see if a quick reset is effective in your environment, and to improve the Reset report. Still, your 0.8.4.1 should work as is. ------------------------------------------------------------------------ POMdev's Profile: http://forums.slimdevices.com/member.php?userid=70558 View this thread: http://forums.slimdevices.com/showthread.php?t=111663 _______________________________________________ Radio mailing list Radio@lists.slimdevices.com http://lists.slimdevices.com/mailman/listinfo/radio