Multiple beacons offline together

siclark · Post by **siclark** » Wed Oct 16, 2019 3:31 am

Hi Karl
I have a mesh of 6 Pis now and all is working well generally, but I have a couple of devices that I use pibeacon to detect presence and automate actions when gone but they keep going down and almost immediately going back on line. I've added them to Grafana to keep an eye on behaviour and see that 2 of them, both beacons, are going down at the same time, which makes me think its an issue with pibeacon, not the devices, especially as each is normally seen by at least 4 Pis.
Any ideas on how to debug?
Screenshot shows them going off together, then appears the car takes a few seconds longer to come back Up at times. The bag is kept in the house, so seen by other Pis to the car.
Thanks

kw123 · Post by **kw123** » Wed Oct 16, 2019 4:09 am

Increase the expiration time for these beacons. In beacon device edit.
That should fix it. May be. Hopefully

Karl
Ps what is is now?

Sent from my iPhone using Tapatalk

siclark · Post by **siclark** » Wed Oct 16, 2019 4:13 am

3 and 4 minutes. I understand how that helps where only one Pi or at edge of detection.
But why would multiple pi's lose connection to 2 different beacons at exactly same time?

Sent from my iPhone using Tapatalk

kw123 · Post by **kw123** » Wed Oct 16, 2019 4:25 am

Anything else in the plugin.log file?
3..4 minutes is pretty long.

What are the exact times when they go down?

The rpis get a parameters update very 15 minutes or so. Depending on circumstances that might restart either master.py or reboot. A restart of master takes a few secs and should not matter

Could you post the file for the rpis
~/pibeacon/permanent.log

That should show when a self inflicted reboot occurs

Sent from my iPhone using Tapatalk

Ps you can get the exact updown times with plugin utilities/sql reports

siclark · Post by **siclark** » Wed Oct 16, 2019 6:04 am

kw123 wrote:Anything else in the plugin.log file?
3..4 minutes is pretty long.

What are the exact times when they go down?

The rpis get a parameters update very 15 minutes or so. Depending on circumstances that might restart either master.py or reboot. A restart of master takes a few secs and should not matter

Could you post the file for the rpis
~/pibeacon/permanent.log

That should show when a self inflicted reboot occurs

Ps you can get the exact updown times with plugin utilities/sql reports

Yes set long to try and avoid this.

Sadly my utilities plugin instance is also not working.

utilities myLog in Line '95' has error='MLX instance has no attribute 'errorLog''
utilities error in querry:
/bin/sh: /Library/PostgreSQL/10/bin/psql: No such file or directory

But looking at Grafana, yes, they appear to be 15 mins apart, 19:39:18 off, then on approx 50 seconds later. then off 15 mins later again etc. I see some constant on times for 30 mins, and then it can be good for 9 hours say.

I cant find a permanent.log on my mac server

siclark · Post by **siclark** » Wed Oct 16, 2019 6:25 am

Ah.. its on the rpis themselves.

I see that both devices went down for all rpis at 05:44 this morning, but no corresponding entry in the log files. I do see lots of these restartMyself entries for all rpis. Ths ones this morning seem to be more like 10 and 20 mins apart.

kw123 · Post by **kw123** » Wed Oct 16, 2019 6:44 am

Are you using the get battery info
Option?

Sent from my iPhone using Tapatalk

siclark · Post by **siclark** » Wed Oct 16, 2019 7:03 am

kw123 wrote:Are you using the get battery info
Option?

Sent from my iPhone using Tapatalk

No

kw123 · Post by **kw123** » Wed Oct 16, 2019 7:05 am

I am traveling right now back home nov 5.

I will try to check. But it might take some time.

Sent from my iPhone using Tapatalk

kw123 · Post by **kw123** » Thu Oct 17, 2019 9:09 am

after looking at the log file, it seems that the BLE stack is not working correctly.

what is your setup?
RPIx
OSv
BLE setup (blé on main board, dongle ...)
could you send one of the parameter files (blank the uid/passwds)
?

Karl

kw123 · Post by **kw123** » Thu Oct 17, 2019 3:08 pm

I fixed the
"utilities myLog in Line '95' has error='MLX instance has no attribute 'errorLog''
in utilities, but not tested:
https://www.dropbox.com/s/dp5avywpjqcir ... n.zip?dl=1

if you could try..?

I need to correlate the exact time (+-3 minutes) of beacons going down and the log file entries

Karl

siclark · Post by **siclark** » Fri Oct 18, 2019 9:09 am

Thanks. That version doesnt work though.

However, in looking at the settings on both the pibeacon devices, and the pis themselves, and populating the drop downs that were not populated, I seem to have done something that has got it working much better now. For last 48 hours I see zero random short blips to down then back up again. All now looks good.

Thanks

kw123 · Post by **kw123** » Sat Oct 19, 2019 2:35 am

I see zero random short blips to down then back up again.

.. that is still not good.

would like to understand how that happens ..

siclark · Post by **siclark** » Sat Oct 19, 2019 12:51 pm

Sorry, that was meant to mean I no longer have the problem. On attached picture, up to morning of 10/16 you can see the small green lines in amongst the red blocks, that were the 10-30 second down times when the devices didnt move. Since then, these are no longer occuring. Only thing I did was tidy up some settings I think.. but all seems to be working now.

kw123 · Post by **kw123** » Sat Oct 19, 2019 2:54 pm

So what settings did you change?

Sent from my iPhone using Tapatalk

Indigo Domotics