Flukso crashing on a daily base

Hello,

since my first installation 6 weeks ago, my device has been "crashing" on a daily base.

Below an extract of the logread:

Jan 1 01:00:56 flukso-13187f daemon.info supd[1442]: Sucessfully started with DEAMON=supd and DAEMON_PATH=/var/run/supd
Jan 1 01:00:56 flukso-13187f daemon.info tmpod[1447]: Sucessfully started with DEAMON=tmpod and DAEMON_PATH=/var/run/tmpod
Jan 1 01:00:56 flukso-13187f user.info sysinit: Daemon returned 0 as return value.
Jan 1 01:00:56 flukso-13187f kern.err kernel: [ 56.710000] uart: rx fifo overflow
Jan 1 01:00:56 flukso-13187f daemon.info kubed[1450]: Sucessfully started with DEAMON=kubed and DAEMON_PATH=/var/run/kubed
Jan 1 01:00:56 flukso-13187f user.info sysinit: Daemon returned 0 as return value.
May 22 14:50:51 flukso-13187f user.info heartbeat[1455]: POST https://api.flukso.net/device/{GUID}: 200
May 22 14:51:24 flukso-13187f authpriv.info dropbear[1474]: Child connection from 10.0.0.1:32966
May 22 14:51:24 flukso-13187f authpriv.info dropbear[1474]: Exit before auth: Exited normally
May 22 14:51:50 flukso-13187f cron.err crond[1218]: time disparity of 27028130 minutes detected
May 22 15:00:01 flukso-13187f cron.info crond[1218]: crond: USER root pid 1486 cmd /usr/bin/heartbeat 0
May 22 15:00:01 flukso-13187f cron.info crond[1218]: crond: USER root pid 1487 cmd /usr/bin/fcheck
May 22 15:00:03 flukso-13187f user.info heartbeat[1489]: POST https://api.flukso.net/device/{GUID}: 200
May 22 15:15:01 flukso-13187f cron.info crond[1218]: crond: USER root pid 1540 cmd /usr/bin/fcheck
May 22 15:21:29 flukso-13187f daemon.err tmpod[1447]: /usr/sbin/tmpod.lua:106: MQTT error: The connection was lost.
May 22 15:30:01 flukso-13187f cron.info crond[1218]: crond: USER root pid 1589 cmd /usr/bin/fcheck
May 22 15:36:40 flukso-13187f daemon.err tmpod[1447]: /usr/sbin/tmpod.lua:106: MQTT error: The client is not currently connected.
May 22 15:39:27 flukso-13187f daemon.err tmpod[1447]: /usr/sbin/tmpod.lua:106: MQTT error: The connection was lost.

Once the "MQTT connection was lost" happens, all data sending stops.
If I notice this in the first 24hours after the problem has happened, I can still ssh into the device and reboot it.
If it's longer than 24h, the device becomes unresponsive. I need to pull the (physical) plug and let it restart like that.

Anyone had a similar experience? The only 'tweek' I did was to open SSH/web for the WAN port (don't have a physical connection available).

Thanks
Danny

gebhardm's picture

This really sounds weird. The MQTT server is "standard software" on the FLM, nothing tweaked. Have you tried a factory reset to bring the overall condition of the base system "back into shape"? That is actually the only option I see.

bitnoise's picture

Hi,

I have done that already once, after I switched to ethernet connection and was unable to connect via utp (I assume you mean the small reset button on the back)?

Thanks
Danny

bitnoise's picture

Ah ok, I'll try it again, because past Monday I have moved the powerclamps to another place in the switchboard and now all my gauges show me for 90% of the load is W- instead of W+ like before, and even Pf - (even going to -2.1).

I'll try it this evening.

Thanks
Danny

bitnoise's picture

Hello,

I've tried multiple times over the weekend to do 'any' kind of reset and nothing works. Not the 'network only' reset, not the full reset. I also came across another post of someone who said all cables had to be removed from the device before doing the reset, but there I have the same issue.

Any other hints? I get the feeling I just have a faulty device.. :(

Thanks
Danny