These are chat archives for nightscout/intend-to-bolus

11th
Mar 2019
Martin Haeberli
@mhaeberli
Mar 11 00:41
In case this wasn’t yet posted above, Enhanced Radio Devices again has Edison-compatible Explorer boards available for order. Deliveries promised April 3 or earlier
On an unrelated note (I -may- try to post the graph later) - our T1D running SMB OpenAPS dev went from 165 at 1:30 AM just before bedtime to 287 at 2:30 AM with NO added carbs; appeared to be flat at 165 before 130, with pretty much no COB, no IOB besides basal. Trying to figure out what model would explain such a rapid rise with no carbs... suggestions welcome
zasteph
@zasteph
Mar 11 01:05
@mhaeberli would it be possible you had an air bubble in the tube? (missing insulin for the duration of the bubble)
Martin Haeberli
@mhaeberli
Mar 11 01:10
@zasteph - interesting hypothesis though previous set change was 2 days before; perhaps a bubble was left in the reservoir...
@scottleibrand - I’m aloft, and also my gitter search skills aren’t the best - I had an interaction with you here maybe a year ago about rsync ing autotune from multiple rigs to the cloud and back so that all the rigs end up on the same page... - if it’s easy, please remind me of the incantations (as I recall, you scheduled same in cron, and so I’d love the cron cheatsheet for it too) ; if not, no worries. Appreciate all you and @danamlewis do
Jon Cluck
@cluckj
Mar 11 01:49
@tzachi-dar yep, if you have the adapter that'll probably work well. I've had pretty good results with a 1/4 wave dipole with the rfm69 breakout board radios
the cost of rfm69 rigs are very low, especially if they're going to be stationary
vsherbeck
@vsherbeck
Mar 11 04:49
Hi everyone; I'm trying to set up my first rig. I have a number of questions so I'll apologize ahead of time. First up: I set up a single wireless connection during my run through PiBakery... how hard is it to add additional connections ?
Or does it involve starting completely over
Scott Leibrand
@scottleibrand
Mar 11 05:05
Easy to add, as long as you can edit the wpa_supplicant.conf config file correctly and don’t introduce any syntax errors.
@mhaeberli our rsync config is complex and fragile, and still occasionally fills up the rigs’ disks. Maybe better to sync through NS?
Martin Haeberli
@mhaeberli
Mar 11 05:14
Thx. Sync through NS? All the rigs refer to the same NS Instance. What other magic needed? (and I’m still willing maybe to try brittle sync...)
vsherbeck
@vsherbeck
Mar 11 05:34
Ok, so I'm now at the point of running the setup--if I want to enable carbsReq Pushover alerts--what or how do I assign a pushover API token?
vsherbeck
@vsherbeck
Mar 11 05:40
Seems like I have to update something in nightscout in conjunction?
ohh I see it's via the pushover app and all
kempera
@kempera
Mar 11 06:23
Dear all, I have used several APSs in parallel and the results were pretty good - much better than with just one machine, which is the reason why I wanted to share some thoughts with you: Perhaps it might make even sense to think about „parallel computation“, i.e. let multiple APSs compete together for the „optimal insulin management path“ and perhaps the number of calculation iterations should depend on the „volatility of blood sugar“, i.e. more calculations if the blood sugar is more volatile...I know that the communication of several machines should ideally be muted once they checked for other machines, but somehow I have the impression that this doesn’t work in my 3 APS installations. Hence, the APS are running „in parallel“ and challenge each
other... which is probably not perfect for the pump and battery, but the short-term results are pretty amazing - which is the reason why I wanted to share my thoughts with you...
Please advise me why this is probably not a good idea and what what similar experimental setups have shown... unfortunately I haven’t found any posts or articles I
on this topic, yet. Thanks again tot he whole community for your support - My family and I highly appreciate it! Thanks, Andreas
tzachi-dar
@tzachi-dar
Mar 11 07:20
@vsherbeck are you talking about adding another wifi connection?
vsherbeck
@vsherbeck
Mar 11 07:22
Yeah. I think I'll be able to puzzle it out with the mention of the wpa config stuff
vsherbeck
@vsherbeck
Mar 11 07:27
Ok so... another question/issue. my rig will be the RFM based; but... I actually got the new product adafruit put out recently with the LED: https://www.adafruit.com/product/4072 . I've been following the set up instructions for the RFM-based set up but now I'm running into some issues as it's not recognizing some of the commands I'm told to put in to "build the right pump communication libraries" per the guide.
Also... it's now displaying "OpenAPS Explorer HAT" with the unicorn on the LED now...
So I guess I'm wondering... do I need to treat this more like a different set up because it has an LED ?
tzachi-dar
@tzachi-dar
Mar 11 08:00
Thanks @cluckj and @ecc1 ordered 4 antennas and 4 rfm69 and cables. Hope to be able to work it out..
viq
@viq
Mar 11 10:36
Random but amusing ;)
python3 ~/Work/Own/oref0-viq/bin/get_profile.py --nightscout https://my.nightscout.url display --name 'OpenAPS Autosync' | tail -n +2 | jq .basal | grep value | cut -d\" -f4 | sparklines
▁▁▁▃▆▅▄▆▇█▆▆▇▇▇▆▆▆▆▆▆▆▆▆
samueldemers2
@samueldemers2
Mar 11 12:12
@davidvjr. I have the same situation as you did. Could you tell me what you did to get through it?
Im talking about the error on march 6
samueldemers2
@samueldemers2
Mar 11 12:53
@davidvjr my bad. The error you got was on march 5 at 21:00.
Scott Leibrand
@scottleibrand
Mar 11 12:58
@mhaeberli the magic of some more development work, to finish making autotune upload the tuned profile to NS, and then make it download it as well.
Jon Cluck
@cluckj
Mar 11 12:59
@vsherbeck I have some beta code that gets all the hardware working located here: https://github.com/cluckj/oref0/tree/dev
Scott Leibrand
@scottleibrand
Mar 11 13:01
@kempera which systems did you have running together? We routinely run multiple OpenAPS rigs at the same time, but that’s for better radio coverage, not for algorithmic diversity. The closest I’ve come to the latter is running OpenAPS rigs with different generations of the same algorithm during testing, running ones with different preferences enabled (usually accidentally), or running off of two different CGMs. I’m not sure that combining the results at the level of “competing to set different temp basals” is ideal, particularly if you enable SMB, as that means that the more aggressive algorithm gets disproportionate say (as it’s possible to deliver a lot more than 2x your normal basal, but not possible to deliver less than zero).
And from a practical perspective, there’s the issue of mediating access to the pump via the radio: OpenAPS rigs are designed to be polite and wait for silence before transmitting, but I’m not aware of any other system designed to do that, so the others tend to interfere quite a bit (if they work at all) if you try to use more than one controller with the same pump.
Raymond Richmond
@PedanticAvenger
Mar 11 15:12
@bud9hb Check the dhclient config files, leases file. I have issues with them getting corrupted once and a while which breaks things. In the case of the leases file just empty it and reboot.
Raymond Richmond
@PedanticAvenger
Mar 11 16:56
Has anyone gotten rsyslog working on an edison with an elasticsearch target? I"m trying to get Humio working as a papertrail alternative and having "package doesn't seem to have everything it needs" issues and would like to confirm I'm not being brain deficient.
zasteph
@zasteph
Mar 11 17:20

Hi, I’m building a new rig, I was at the bootstrap script.
I ran it once, it stopped asking me to change the root passw from edison, after installing things for a while.
I changed the passw and ran again bash /tmp/openaps-install.sh, and after a few pages of “… is already the newest version” & “Requirement already up-to-date” it stopped with a bunch of errors in Cleanup:
Cleaning up...
installed /etc/udev/rules.d/80-medtronic-carelink.rules
installed /etc/udev/rules.d/80-dexcom.rules
Installing bash completion script /etc/bash_completion.d/python-argcomplete.sh
/usr/local/bin/json -> /usr/local/lib/node_modules/json/lib/json.js
npm WARN deprecated crypto@0.0.3: This package is no longer supported. It's now a built-in Node module. If you've depended on crypto, you should switch to the one that's built-in.
npm WARN engine request@2.88.0: wanted: {"node":">= 4"} (current: {"node":"0.10.29","npm":"1.4.21"})
npm WARN engine form-data@2.3.3: wanted: {"node":">= 0.12"} (current: {"node":"0.10.29","npm":"1.4.21"})
npm WARN engine har-validator@5.1.3: wanted: {"node":">=6"} (current: {"node":"0.10.29","npm":"1.4.21"})
npm WARN deprecated node-uuid@1.4.8: Use uuid module instead
npm WARN engine tough-cookie@3.0.1: wanted: {"node":">=6"} (current: {"node":"0.10.29","npm":"1.4.21"})
npm WARN engine har-schema@2.0.0: wanted: {"node":">=4"} (current: {"node":"0.10.29","npm":"1.4.21"})
npm ERR! cb() never called!
npm ERR! not ok code 0

Should I run it again? Run an “uninstall”?
Thank you for your help.

Scott Leibrand
@scottleibrand
Mar 11 17:22
try running it again a few times. it seems that cb() never called is related to a transient issue with npm, possibly due to something like number of dependent packages vs. resources on the Pi.
zasteph
@zasteph
Mar 11 17:24
@scottleibrand , thank you! You mean just "bash /tmp/openaps-install.sh", or I need this line too: "curl https://raw.githubusercontent.com/openaps/oref0/master/bin/openaps-install.sh > /tmp/openaps-install.sh"
Scott Leibrand
@scottleibrand
Mar 11 17:31
the latter just downloads the script. once that's done, re-running it just requires the former (bash /tmp/openaps-install.sh)
zasteph
@zasteph
Mar 11 17:32
@scottleibrand I'll try that. Thank you!
TranceCake
@TranceCake
Mar 11 18:45
is there a clean way to uninstall open aps and try again or do I need to re-flash my sd card?
my second rig is crashing ns
Scott Leibrand
@scottleibrand
Mar 11 18:46
only fully clean way is to re-flash
if you think it's something config related you could just delete ~/myopenaps and reinstall
TranceCake
@TranceCake
Mar 11 18:47
I don't really know
<!DOCTYPE html>
     <html>
       <head>
         <meta name="viewport" content="width=device-width, initial-scale=1">
         <meta charset="utf-8">
         <title>Application Error</title>
         <style media="screen">
           html,body,iframe {
             margin: 0;
             padding: 0;
           }
           html,body {
             height: 100%;
             overflow: hidden;
           }
           iframe {
             width: 100%;
             height: 100%;
             border: 0;
           }
         </style>
       </head>
       <body>
         <iframe src="//www.herokucdn.com/error-pages/application-error.html"></iframe>
       </body>
     </html> (NOT VALID JSON: parse error: Invalid numeric literal at line 1, column 10)
 Completed oref0-ns-loop at Thu 7 Mar 10:23:38 CET 2019
this is from ns-loop
but I'm not sure how I can see the request or response that is causing this
looks like it is an error in one of the rigs requests to nightscout
Scott Leibrand
@scottleibrand
Mar 11 18:55
only this one rig is doing it? I'd just reflash before investigating too much
TranceCake
@TranceCake
Mar 11 18:58
yeah only the one rig
kinda hesitant because it took so long to make it work in the first place :)
viq
@viq
Mar 11 19:17
@PedanticAvenger I don't have a Pi, but I have some linux experience, so you could throw some questions my way and I'll try to have a look when I have time
Jon Cluck
@cluckj
Mar 11 19:21
trance, that looks like your NS crashed?
Raymond Richmond
@PedanticAvenger
Mar 11 19:34
@viq You mean edison? if that is what you have (same as myself) try to install the rsyslog package from the master repository (the one in the debian tree is too old) and see if the elasticsearch module install actuall installs the .so lib
Jan Schenk
@jansche
Mar 11 19:45
I'm also still having issues installing dev on an Edison. It doesn't matter if I go from master to dev, or if I Ctrl-C and try to jump to dev directly. It breaks with this:
grafik.png
This is a super clean install with flashing a fresh jubilinux 0.2, every step exactly according to the docs.
Jan Schenk
@jansche
Mar 11 19:50
I'm now installing master. This worked the first time, but the rig was getting borken when trying to switch to dev. Are there any downsides or known problems running dev branch rigs in parallel to master branch rigs?
Scott Leibrand
@scottleibrand
Mar 11 20:07
no, that should be fine. the only difference I can think of is that dev autotune categorizes UAM as basal when there is >1h of carb absorption over the day, so their autotuned profiles may get slightly out of sync eventually
Jan Schenk
@jansche
Mar 11 20:12
ok, thanks, @scottleibrand.
Any idea why the setup fails on dev on Edison? I thought many others are having issues with Pi0s, but it could be this is something more general not platform specific. I saw the nightscout crud error on both, Pi and Edison. And I thought the regarding commit was reverted...
As I now have a spare Edison rig, I can do some bug tracking.
Jon Cluck
@cluckj
Mar 11 20:17
I think that bug was fixed last night, or the night before?
@jansche this is a pull from today?
viq
@viq
Mar 11 20:22
@PedanticAvenger no, I don't have any kind of ARM, I'm on androidaps and just hanging out here ;)
viq
@viq
Mar 11 20:41
@PedanticAvenger apparently also syslog-ng has support for elasticsearch output. Also both apparently can send GELF which humio claims to support, so that may be worth checking out as well
Jan Schenk
@jansche
Mar 11 20:47
@cluckj yes, today's pull. Like three times freshly installed, and running into the error over and over.
Jan Schenk
@jansche
Mar 11 20:59
One seems also not to be able to go back from dev to master. Is that a known issue?
npm run global-install moans about broken packages...
Jon Cluck
@cluckj
Mar 11 21:03
I haven't even tried to go back to master from dev :laughing:
I guess it's known now? :)
(getting my long-lost edison rig updated btw)
Eric
@ecc1
Mar 11 21:05
that's the "burn the ships" strategy :-)
Jon Cluck
@cluckj
Mar 11 21:05
LOL
I know a couple folks updated to the latest dev on the pi today without issue, so I think it's an edison thing
@jansche what's your jq --version?
Jon Cluck
@cluckj
Mar 11 21:13
it should be jq-1.5-1-a5b5cbe
if it's older than that, updating it will probably fix your issue
TranceCake
@TranceCake
Mar 11 21:17
@cluckj yeah it did indeed, re flashing the pi now
Jon Cluck
@cluckj
Mar 11 21:17
@TranceCake you'll still have to fix nightscout :P
TranceCake
@TranceCake
Mar 11 21:18
oh yeah that error was couple of days ago, I redeployed it
Jon Cluck
@cluckj
Mar 11 21:19
okay :)
Jan Schenk
@jansche
Mar 11 21:24
@cluckj apt-get upgrade jq? That thorws an error like that:
Reading package lists... Done
Building dependency tree       
Reading state information... Done
Calculating upgrade... Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
 jq : Depends: libjq1 (= 1.5+dfsg-1.3~bpo8+1) but it is not going to be installed
      Depends: libonig2 (>= 5.9.5) but it is not installable
E: Broken packages
TranceCake
@TranceCake
Mar 11 21:25
oh dear, this is the second time I get a python exception during the installation
Jan Schenk
@jansche
Mar 11 21:25
And yes, I have
root@boon:~# jq --version
jq-1.4-1-e73951f
Jon Cluck
@cluckj
Mar 11 21:25
apt-get -t jessie-backports install jq
TranceCake
@TranceCake
Mar 11 21:25
Exception:
Traceback (most recent call last):
  File "/usr/lib/python2.7/dist-packages/pip/basecommand.py", line 215, in main
    status = self.run(options, args)
  File "/usr/lib/python2.7/dist-packages/pip/commands/install.py", line 353, in run
    wb.build(autobuilding=True)
  File "/usr/lib/python2.7/dist-packages/pip/wheel.py", line 749, in build
    self.requirement_set.prepare_files(self.finder)
  File "/usr/lib/python2.7/dist-packages/pip/req/req_set.py", line 380, in prepare_files
    ignore_dependencies=self.ignore_dependencies))
  File "/usr/lib/python2.7/dist-packages/pip/req/req_set.py", line 620, in _prepare_file
    session=self.session, hashes=hashes)
  File "/usr/lib/python2.7/dist-packages/pip/download.py", line 821, in unpack_url
    hashes=hashes
  File "/usr/lib/python2.7/dist-packages/pip/download.py", line 659, in unpack_http_url
    hashes)
  File "/usr/lib/python2.7/dist-packages/pip/download.py", line 853, in _download_http_url
    stream=True,
  File "/usr/share/python-wheels/requests-2.12.4-py2.py3-none-any.whl/requests/sessions.py", line 501, in get
    return self.request('GET', url, **kwargs)
  File "/usr/lib/python2.7/dist-packages/pip/download.py", line 386, in request
    return super(PipSession, self).request(method, url, *args, **kwargs)
  File "/usr/share/python-wheels/requests-2.12.4-py2.py3-none-any.whl/requests/sessions.py", line 488, in request
    resp = self.send(prep, **send_kwargs)
  File "/usr/share/python-wheels/requests-2.12.4-py2.py3-none-any.whl/requests/sessions.py", line 609, in send
    r = adapter.send(request, **kwargs)
  File "/usr/share/python-wheels/CacheControl-0.11.7-py2.py3-none-any.whl/cachecontrol/adapter.py", line 47, in send
    resp = super(CacheControlAdapter, self).send(request, **kw)
  File "/usr/share/python-wheels/requests-2.12.4-py2.py3-none-any.whl/requests/adapters.py", line 423, in send
    timeout=timeout
  File "/usr/share/python-wheels/urllib3-1.19.1-py2.py3-none-any.whl/urllib3/connectionpool.py", line 643, in urlopen
    _stacktrace=sys.exc_info()[2])
  File "/usr/share/python-wheels/urllib3-1.19.1-py2.py3-none-any.whl/urllib3/util/retry.py", line 315, in increment
    total -= 1
TypeError: unsupported operand type(s) for -=: 'Retry' and 'int'
Jan Schenk
@jansche
Mar 11 21:26
@cluckj same:
The following packages have unmet dependencies:
 jq : Depends: libjq1 (= 1.5+dfsg-1.3~bpo8+1) but it is not going to be installed
      Depends: libonig2 (>= 5.9.5) but it is not installable
E: Unable to correct problems, you have held broken packages.
Jon Cluck
@cluckj
Mar 11 21:26
hm :\
matrixbot
@matrixbot
Mar 11 21:26
@viq:matrix.org Time for dist-upgrade?
@viq:matrix.org Also, is this debian 8 or 9?
Jon Cluck
@cluckj
Mar 11 21:30
@jansche cat /etc/debian_version?
you said jubilinux .2 earlier?
Jan Schenk
@jansche
Mar 11 21:30
8.11
jubilinux 0.2.0
Jon Cluck
@cluckj
Mar 11 21:31
can you try to install those jq libs?
Jan Schenk
@jansche
Mar 11 21:32
I did, libjq1 depends on libonig2, which has no install candidate
how could I get libonig2 installed?
matrixbot
@matrixbot
Mar 11 21:33
@viq:matrix.org apt-cache policy libonig2 IIRC
Jon Cluck
@cluckj
Mar 11 21:34
yes, what viq's ghost said :laughing:
Jan Schenk
@jansche
Mar 11 21:34
root@boon:~# apt-cache policy libonig2
libonig2:
  Installed: (none)
  Candidate: (none)
  Version table:
Jon Cluck
@cluckj
Mar 11 21:35
apt-get update you may have a busted cache
Jan Schenk
@jansche
Mar 11 21:35
running it
Jon Cluck
@cluckj
Mar 11 21:35
libonig2:
  Installed: 5.9.5-3.2+deb8u1
  Candidate: 5.9.5-3.2+deb8u1
  Version table:
 *** 5.9.5-3.2+deb8u1 0
        500 http://http.debian.net/debian/ jessie/main i386 Packages
        100 /var/lib/dpkg/status
that's what I've got
matrixbot
@matrixbot
Mar 11 21:35
@viq:matrix.org apt-cache showpkg libonig2 for more info, but update is a good idea indeed
Jan Schenk
@jansche
Mar 11 21:37
busted cache it was
Jon Cluck
@cluckj
Mar 11 21:37
:thumbsup:
it probably reported that error installing jq waaay up in the install logs
Jan Schenk
@jansche
Mar 11 21:38
Will try to switch to dev now...
Thanks to the both of you! @cluckj && @viq
TranceCake
@TranceCake
Mar 11 21:40
I think it got past the error this time
what could cause this kind of error?
TranceCake
@TranceCake
Mar 11 21:53
@cluckj I got the npm error with cb() again and I did the ulimit thing and now it seems to pass, so that might be a solution after all
yup it passed
Jon Cluck
@cluckj
Mar 11 22:13
running it a second time also works so :shrug:
TranceCake
@TranceCake
Mar 11 22:41
running into the
-bash: go: command not found
thing again
even after manually installing go as per the instruction ecc1 gave me few days ago
/usr/local/go is there
it just won't run
aparently the pi needs a reboot after the go installation or something cause that fixed it