nightscout/intend-to-bolus

Discussions around DIY closed looping with #OpenAPS

tynbendad

@tynbendad

about the radio hack that attempts to guess your pump's serial number - has anyone tried to detect a hack in progress (e.g., from an explorer or explorer hat rig)?

is it feasible?

Eric

@ecc1

@tynbendad not sure what hack you mean, but it's trivial to snoop on pump comms in progress and extract the serial number, since it's not encrypted in any way. Impossible to detect, since it's receive-only

tynbendad

@tynbendad

Oh, ok that's worse than i thought. Thought they had to try every serial number, but maybe that's just for pumps that aren't using comms.

Kelly Watson

@kellyisworking

https://gofundme.com/f/support-katie-disimone-and-her-family

jaybaron1

@jaybaron1

Thanks for sharing.

Dave Acklam

@dcacklam

@tzachi-dar rpi4 works as a rig but the battery circuitry on the explorer hat is too weak to power it.

@tzachi-dar So only really good for a desktop rig.

Use a pi3A+ or pizero for portable

tzachi-dar

@tzachi-dar

Thanks @dcacklam is using pi3A+ with a battery a reasonable thing to do? Seems like it's power usage is relatively high.

Dave Acklam

@dcacklam

@tzachi-dar just use a 6000mah battery

Caleb Seekell

@jauntySerpent

My backup rig isn't working at the moment so I'm unable to test a completely fresh install, but there appears to be issues with the current master branch of oref0 (0.7.0) and token-based NS authentication. I'm finally getting around to updating my rig with token auth instead of api_secret, but I've been troubleshooting an issue for the past several weeks where the rig does not include the token in several GET requests, resulting in an permissions error and prevents the rig from downloading BG data.

Jul 21 07:00:07 ********* heroku/router at=info method=GET path="/api/v1/entries/sgv.json?count=288" host=*********.herokuapp.com request_id=a57e9b7b-cd93-4c10-8903-f73eb9f70d52 fwd="*********" dyno=web.1 connect=0ms service=5ms status=401 bytes=560 protocol=https
Jul 21 07:00:07 ********* app/web.1 Resolving secret/token to permissions failed
Jul 21 07:00:12 ********* heroku/router at=info method=GET path="/api/v1/entries/sgv.json?count=12" host=*********.herokuapp.com request_id=ee5eef71-627b-4f36-81fd-2eefeb116d9b fwd="*********" dyno=web.1 connect=0ms service=473ms status=401 bytes=560 protocol=https
Jul 21 07:00:12 ********* app/web.1 Resolving secret/token to permissions failed

The current version of dev does not have this problem, and I finally found this commit which seems to describe the issue perfectly openaps/oref0@89a90b8 . Can anyone else confirm whether they are seeing this issue?

Dana Lewis

@danamlewis

I am not sure if that’s the only thing broken in master but you’re not the only one who needed to switch to dev recently and found it resolved the/an issue.

If we can get some volunteers to test some fresh installs and report back (here or in the open PR of dev to master), we could do a release (dev to master) and get master back to a working state.

old-square-eyes

@old-square-eyes

How do I disable Autotune, but leave it to run each night on my rig?

I just want to manually observe recommendations

I keep dropping low and it keeps recommending more insulin. I can't figure out why.

I'm entering carbs.

Constantly getting negative IOB

old-square-eyes

@old-square-eyes

Do I just change "enable": " autotune ", to "enable": "",?

In prefs

Caleb Seekell

@jauntySerpent

@danamlewis I am happy to test as soon as I get my backup rig running.

old-square-eyes

@old-square-eyes

Or do I delete the whole line

refields

@refields

For a few days now there has been no activity on my nightscout

I didnt change anything in nightscout settings, Dexcom Share is on with connectivity, and Heroku plug size is
58.2 MiB of 500 MiB. Anyone know what else I can try or check?

Eric

@ecc1

@refields have you checked if your mongodb is full?

is your openaps successfully uploading to NS? what do the logs say?

refields

@refields

@ecc1 it looks like my mongodb is full

how do I clear it to zero?

refields

@refields

openaps is creating new logs in paper trail but not looping

Most recent logs:

Eric

@ecc1

@refields see https://openaps.readthedocs.io/en/latest/docs/Troubleshooting/Rig-NS-communications-troubleshooting.html#mlab-maintenance

refields

@refields

@ecc1 I dont think those instructions apply anymore since the mLab migration to Atlas which happened in November 2020. Are there instructions for doing this in the Atlas UI?

Eric

@ecc1

@refields I found this using Google; you can too. I don't use a hosted mongo service, so YMMV: https://nightscout.github.io/troubleshoot/troublehoot/#database-full

refields

@refields

@ecc1 this worked. thank you again. I promise I did try Google initially, and unsuccessfully. I think I even had to do this once before. The problem is life goes on fine for 6-12 months with the CGM-rig working properly and then something goes awry and I forget all the steps. Really appreciate it.

Eric

@ecc1

@refields glad it worked. I have a repeating 6 month reminder on my calendar to do this, btw

tzachi-dar

@tzachi-dar

@ecc1 @refields openaps/oref0#1232 has a script that deletes old NS entries every night. Actually it does not delete them but rather copy them locally. Fill free to use it.

Foxy7

@Foxy7

@kellyisworking thanks for the link.

refields

@refields

@ecc1 thats a great idea will do the same. Thank you @tzachi-dar

Milan

@milanlakhani

Hi all :) we are trying to collect tracing data on an OpenAPS Raspberry Pi rig for our voluntary safety project, we were wondering if anyone here would have time to send us some trace logs of their rig while their loop is running if we send them instructions on how to do this? Also if anyone would like to look at our safety analysis so far? :)

garetis

@garetis

So I bought a Pi4 to upgrade from the Pi0. The PiHAT doesn't seem to fit as nicely onto the 4 as the zero. I looked through the docs. I don't see anything that explicitly says it won't work, but is that a compatible setup or do I need to go and get another bonnet instead?

Sarah Park

@iamgopark

Hi, I have been trying to install OpenAPS on Pi Hat, Medtronic 722 (which serves as CGM and Pump) and got stuck in error messages for several weeks, Profile invalid, Couldn't refresh_old_profile, etc. The signal between Pi Hat and Pump seems to be strong, and the setup retrieves the data OK and updates Nightscout with BG, but it does not generate any treatment files (no files within enact). I followed Dave Cole's note (http://www.dc9.com/openapsonrpizero/InstallingOpenAPSMasteronRPIZeroW.html). I also tried to install the Raspberry Pi Desktop version, but nothing seems to work. Any suggestions would be appreciated.

'code'
Starting oref0-pump-loop at Sun 1 Aug 03:54:07 CDT 2021 with 2 second wait_for_silence:
MDT CGM configured; not waiting
Listening for 2s: .No interfering pump comms detected from other rigs (this is a good thing!)
Continuing oref0-pump-loop at Sun 1 Aug 03:54:11 CDT 2021
Preflight OK.
Attempting to retrieve MDT CGM data from pump
MDT CGM data retrieved
Old settings: pump profile refresh unsuccessful; trying again on next loop
Invalid carb_ratios.json
Profile invalid: -rw-r--r-- 1 root root 596 Aug 1 02:55 settings/profile.json
No more radio errors; canceling reboot
pump profile refresh unsuccessful; trying again on next loop
Couldn't refresh_old_profile
oref0-pump-loop failed. pump_loop_completed more than 15m old; waiting for 23 s silence before mmtuning
'code'

Foxy7

@Foxy7

@iamgopark that looks to me like it's not speaking to the pump

try this section

Where communities thrive

People

Repo info

Activity