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

19th
Sep 2017
Eric
@ecc1
Sep 19 2017 00:25

I just spent a long time finding out why some of my Dexcom readings weren't being uploaded correctly to Nightscout, even though they were present on the Dexcom receiver (my data source). It turned out that through clock drift, my G4 sensor had lined up exactly with the 5 minute interval my loop was running at. So I was trying to read the latest EGV and sensor values from the receiver right as they were being received.

I expected that the sensor value would arrive, and then the receiver might take a few seconds to produce and store the EGV record (that's what their timestamps indicate). But something more than that was going on: sometimes the sensor value wouldn't be available even though the EGV value was.

I'm curious whether anyone sees something like this with openaps rigs. FWIW, my fix was simply to "wait until next time" if the most recent reading didn't have both cooked and raw data available.

Scott Leibrand
@scottleibrand
Sep 19 2017 00:43
I've never noticed that, probably because oref0 doesn't do things on that regular of a schedule.
Not sure if the built in retires would do so in that case or not, but we also get data from NS (uploaded via Dexcom Share), so that would still work.
Chris Reilly
@rainbow-lazer
Sep 19 2017 00:47

Hello, I'm attempting to run autotune locally on OSX. I've done this in the past just fine, but hitting some snags tonight.

I keep getting the error message ERROR: API_SECRET is not set when calling oref0-autotune.sh
So it looks like that is not defined as an environmental variable. Weird thing is, I've never had that issue in previous local runs.

So far, I've tried adding it with the Terminal command $ export API_SECRET=token=name-1123jh12k3jh1k23h as well as manually adding that to my ~/.bash_profile file. Both techniques give me this error:

/usr/local/bin/oref0-autotune: line 150: json: command not found
Grabbing NIGHTSCOUT treatments.json for date range...
Query: https://url.herokuapp.com/find%5Bcreated_at%5D%5B%24gte%5D=2017-09-15T20:00-04:00&find%5Bcreated_at%5D%5B%24lte%5D=2017-09-18T00:00-04:00
/usr/local/bin/ns-get: line 29: ${API_SECRET,,}: bad substitution
/usr/local/bin/ns-get: line 40: ${API_SECRET,,}: bad substitution
Couldn't download ns-treatments.json

Any help is much appreciated!

garykidd
@garykidd
Sep 19 2017 00:57
Thumbs up for oref1. Really working well for me. FYI.
Dana Lewis
@danamlewis
Sep 19 2017 02:45
@rainbow-lazer pop into the Autotune channel - someone had same issue this am and noted his fix there
Chris Reilly
@rainbow-lazer
Sep 19 2017 02:56
@danamlewis gotcha, thanks!
Thacher
@thacherT1D
Sep 19 2017 03:10
Maybe a silly question -- should I have a second rig to run offline?
Or use the same one and make settings updates per the documentation
Scott Leibrand
@scottleibrand
Sep 19 2017 03:23
Yes, we highly recommend two rigs for backup and so you can upgrade one at a time and get everything the way you want without time pressure.
GreatEscape
@GreatEscape
Sep 19 2017 04:19
image.png
@scottleibrand , @Kdisimone - I followed the directions in the docs and installed the decocare fix a second time. I changed the carb ratio to 26. When autotune ran at 12:05, it changed 26 to 0.26 and this resulted in a CSF of 581.538. Our ISF is usually 21-23 and wouldn't normally be impacted by this. But autotune has been tuning it above 25 and I wonder if background calculations may be somehow impacted by this. I have a Canadian 754 pump in mmol - not sure if that might be a factor in this.
Cylon75
@Cylon75
Sep 19 2017 07:07

@danamlewis
My timezone is Europe/Amsterdam so that should be good. If have found out that I can see the IFTTT button action in the "run console" on the heroku site. Restarted my dyno and it seems to work now somehow.I still have those future treatments or is that normal?

@Kdisimone
After restarting I was able to verify that I need to enter those values in mg/dl. Thanks.

Dana Lewis
@danamlewis
Sep 19 2017 07:21
Try deleting the future treatments. You shouldn't have those.
Cylon75
@Cylon75
Sep 19 2017 07:28
@danamlewis
I have deleted those but in like 60 seconds they are back. It seems to be last two hours and we are UTC/GMT +2 here in the Netherlands. Does that have something to do with it? My pump,CGM and rig are set at the same time.
PieterGit
@PieterGit
Sep 19 2017 10:13
@Cylon75 what kind of setup do you use. Dexcom or Medtronic CGM. What uploaded so you use?
Uploader
Cylon75
@Cylon75
Sep 19 2017 10:13
I use a Dexcom G4 without share and is directly connected to my rig
garykidd
@garykidd
Sep 19 2017 12:33
i accidentally entered rescue carbs instead of 36 entered 366...can i eliminate that somehow?
David
@davidkeddydb
Sep 19 2017 13:52
@garykidd just go into NS/reports/treatments and delete it or edit it there
David
@davidkeddydb
Sep 19 2017 13:59
I can just redo set up but is anyone else getting this error working with 5.3dev, fatal: loose object? Starting supermicrobolus pump-loop at Tue Sep 19 09:50:02 EDT 2017 with 4 second wait_for_silence:
Waiting up to 4 minutes for new BG: glucose.json newer than pump_loop_completed
Radio ok. Listening: .No pump comms detected from other rigs
fatal: loose object b5eb8038fe4136ed62fadb39ac7dcb016f969a3e (stored in .git/objects/b5/eb8038fe4136ed62fadb39ac7dcb016f969a3e) is corrupt'
SMB pump-loop failed. grep: monitor/temp_basal.json: No such file or directory
pump_loop_completed more than 15m old; waiting for 30s silence before mmtuning
Radio ok. Listening: .No pump comms detected from other rigs
Listening for 30s silence before mmtuning: ...No pump comms detected from other rigs
mmtune: fatal: loose object b5eb8038fe4136ed62fadb39ac7dcb016f969a3e (stored in .git/objects/b5/eb8038fe4136ed62fadb39ac7dcb016f969a3e) is corrupt'
"916.540", 5, -52 Unsuccessful supermicrobolus pump-loop at Tue Sep 19 09:52:06 EDT 2017
David
@davidkeddydb
Sep 19 2017 14:23
Well, reinstall won't work as I need to get the .git object out first. Any ideas how I can get rid of the .git? Otherwise I will just delete the directory and reinstall. Dont know what happened. Rig was working all day yesterday without issue Checking /root/myopenaps: Reinitialized existing Git repository in /root/myopenaps/.git/
error: object file .git/objects/b5/eb8038fe4136ed62fadb39ac7dcb016f969a3e is empty
error: object file .git/objects/b5/eb8038fe4136ed62fadb39ac7dcb016f969a3e is empty
fatal: loose object b5eb8038fe4136ed62fadb39ac7dcb016f969a3e (stored in .git/objects/b5/eb8038fe4136ed62fadb39ac7dcb016f969a3e) is corrupt
Can't init /root/myopenaps
Elisakelley
@Elisakelley
Sep 19 2017 15:06
@danamlewis @scottleibrand I am running openAPS on an iphone 7. I've used BT tethering for a while with no problems until the other day when it just stopped working. the blue bar comes up on my phone showing a connection but the rig can't seem to get an IP address. I've tried using 2 different phones with BT tethering and neither gives me an inet address after running ifconfig bnep0. I'm able to connect to my phone's hotspot via BT with an iPad and my Windows laptop. not sure what else I can do, any ideas?
I've re-ran the setup script and done everything in the docs
David
@davidkeddydb
Sep 19 2017 15:13
@Elisakelley Are you able to connect to the phone via it as a hotspot?
kallnap
@kallnap
Sep 19 2017 15:13
@savek-cc I thought the API was open for developpers before
Elisakelley
@Elisakelley
Sep 19 2017 15:14
@davidkeddydb using wifi as opposed to BT? yes
David
@davidkeddydb
Sep 19 2017 15:14
ok
what happens when you go through the BT pairing process? You have it as a paired-device is just won't connect?
Elisakelley
@Elisakelley
Sep 19 2017 15:16
yep, running sudo bt-pan client AA:BB:CC:DD:EE:FF establishes a connection and I see the blue bar on the phone saying the rig is connected to the hotspot
but sudo dhclient bnep0 and ifconfig bnep0 don't show any inet address
David
@davidkeddydb
Sep 19 2017 15:18
Double check nothing is wacky in your /etc/network interface file
Meghan Rutledge
@mar5477
Sep 19 2017 15:23
image.png
I bolused this morning at 6:30 for breakfast while I wasn't connected to my rig. Shouldn't it have pulled that in when it connected at 8:30?
Scott Leibrand
@scottleibrand
Sep 19 2017 15:37
@davidkeddydb you can just rm -rf ~/myopenaps/.git/ to delete your existing git repo, and then re-run oref0-setup
@mar5477 yes, it should pick up boluses from the pump for the last 24h. I'm not sure if there are any race conditions in the upload-to-NS part, though. does the IOB from when it set the first zero temp reflect the previous bolus?
Elisakelley
@Elisakelley
Sep 19 2017 15:39

@davidkeddydb

interfaces(5) file used by ifup(8) and ifdown(8)

auto lo
iface lo inet loopback

auto usb0
iface usb0 inet static
address 192.168.2.15
netmask 255.255.255.0

auto wlan0
iface wlan0 inet dhcp
wpa-conf /etc/wpa_supplicant/wpa_supplicant.conf

# For WPA
#wpa-ssid Emutex
#wpa-psk passphrase
# For WEP
#wireless-essid Emutex
#wireless-mode Managed
#wireless-key s:password

And the following 4 lines are for when using hostapd...

auto wlan0

iface wlan0 inet static

address 192.168.42.1

netmask 255.255.255.0

can you look at that and see anything weird?

David
@davidkeddydb
Sep 19 2017 15:41
very similar to mine, but here is mine

auto lo
iface lo inet loopback

auto usb0
iface usb0 inet static
address 10.11.12.13
netmask 255.255.255.0

auto wlan0
iface wlan0 inet dhcp
wpa-conf /etc/wpa_supplicant/wpa_supplicant.conf

People who know more than me can likely comment but when I have had this issue, it is usually something got weird in the network/interface file somehow
Scott Leibrand
@scottleibrand
Sep 19 2017 15:44
I don't see any problems there (unless your hotspot happens to use 192.168.2.x)
you could always try re-installing from scratch (including a re-flash) if you can't figure it out.
Meghan Rutledge
@mar5477
Sep 19 2017 15:46
It looks like it. There's 8 COB and .5 U IOB.
image.png
Scott Leibrand
@scottleibrand
Sep 19 2017 15:48
ok, so just an NS-upload sequencing issue then: good. I wouldn't worry about it unless it happens repeatedly and gets annoying, in which case we could look at modifying the code
David
@davidkeddydb
Sep 19 2017 15:48
@scottleibrand You know way more about this than me, but if @Elisakelley put something in on the iface usb0 inet static won't that force it to that address assuming it can even connect?
Meghan Rutledge
@mar5477
Sep 19 2017 15:49
@scottleibrand cool, thanks!
Scott Leibrand
@scottleibrand
Sep 19 2017 15:52
@davidkeddydb the usb0 address is only used for talking to any computer that's connected over the USB OTG port (so you can ssh in that way). however, that address is always active whether anything is plugged in or not, so having it overlap with the addresses that any other interfaces are trying to use causes conflicts
David
@davidkeddydb
Sep 19 2017 16:00
@scottleibrand Whether @Elisakelley got anything out of that, I learned something for the day.
Scott Leibrand
@scottleibrand
Sep 19 2017 16:05
it's unlikely to be her issue: just a clue that this rig was set up manually, not with openaps-bootstrap
(the start-from-scratch copy and paste code block that's now in the docs)
Meghan Rutledge
@mar5477
Sep 19 2017 16:13
okay, @scottleibrand another weird thing. My COB which was decaying as expected suddenly jumped up even though there wasn't anything entered
image.png
Elisakelley
@Elisakelley
Sep 19 2017 16:32
@scottleibrand should I re-install from scratch then using the openaps-bootstrap?
Scott Leibrand
@scottleibrand
Sep 19 2017 16:32
@Elisakelley that is one option if you have the time to re-do everything.
Elisakelley
@Elisakelley
Sep 19 2017 16:33
@scottleibrand any other troubleshooting I could do before resorting to that?
Scott Leibrand
@scottleibrand
Sep 19 2017 16:33
not sure
Bluetooth is a bit of a black art: we couldn't get it working for months, and finally found a process that works if we follow it exactly (which is what is now in the docs)
@mar5477 usually that is due to old boluses getting expired before old stacked carbs.
katie disimone
@Kdisimone
Sep 19 2017 16:43
@Elisakelley have you tried a hard reset of your iPhone?
garykidd
@garykidd
Sep 19 2017 16:44
@scottleibrand did you ever contemplate requiring that current BGs must be higher than a certain threshold before any SMBs are adminsitered...had a situation where I was getting SMBs at very low BG levels
katie disimone
@Kdisimone
Sep 19 2017 16:44
@Elisakelley although, I agree, it sounds like a rig-side issue since the other devices catch internet fine on the hotspot still
Scott Leibrand
@scottleibrand
Sep 19 2017 16:51
@garykidd yes, we just had that same issue last night and decided to implement an SMB minPredBG guard next.
garykidd
@garykidd
Sep 19 2017 16:53
awesome! I am comforted by that response.
garykidd
@garykidd
Sep 19 2017 17:04
and just following up....I'm guessing that SMB minPredBG would be a prospective variable? Because, if so, I guess it's a slightly different approach to what I think is the same observed issue.
I was just thinking if BG < [min BG] and/or if BG is expected to fall below [min BG] ....cancel and SMB.
Scott Leibrand
@scottleibrand
Sep 19 2017 17:32
if any COB-predicted BG falls below the cutoff used to avoid high-temps (~75 or ~80 mg/dL for targets in the 90-100 range) then it would wait to SMB.
not sure what you mean by "cancel and SMB"
Nalla306
@Nalla306
Sep 19 2017 17:44
Hi, hope this is the right place to ask. I just installed openaps onto an edison with explorer board and it all went well and was showing up as looping in nightscout along with the pump status. I then went on to the next stage of checking logs by pressing 'l' It gave 'could not get subg_rfspy star
state or version....and its not in an endless reboot
Scott Leibrand
@scottleibrand
Sep 19 2017 17:46
are the two LEDs beside the microUSB ports on?
Nalla306
@Nalla306
Sep 19 2017 17:47
one is on
Scott Leibrand
@scottleibrand
Sep 19 2017 17:47
if it's on solid, that usually indicates a problem with your cc1110 firmware. search the docs for ccprog and follow those steps to re-flash the cc1110
also check that the Edison is seated securely on the Explorer board
and power-cycle everything with the little black button (or by unplugging the battery)
Nalla306
@Nalla306
Sep 19 2017 17:49
okies....sounds tricky, its strange as i bought two edisons and explorers and both have done the same....they connect to nightscout and then when i press l to check logs it errors on both, one was for family member. Will have a read of the ccprog docs thanks
Scott Leibrand
@scottleibrand
Sep 19 2017 17:50
pressing l to view the logs doesn't do anything except display the contents of a text file, so that's not triggering whatever it is.
Nalla306
@Nalla306
Sep 19 2017 17:57
it automatically starts to reboot to fix radio errors before i get chance to really do anything. I'm new and a little slow!
Elisakelley
@Elisakelley
Sep 19 2017 17:57
@Kdisimone i've tried hard reset and resetting just about everything else too :(
Nalla306
@Nalla306
Sep 19 2017 17:59
when it boots... i get one red Error: failed to start hostname service. Not sure if thats relevant
Elisakelley
@Elisakelley
Sep 19 2017 18:02
@Kdisimone if I do wind up re-installing, do I have to do something before starting the install steps in the docs?
as in do I have to clear it out?
David
@davidkeddydb
Sep 19 2017 18:11
@Elisakelley If you use the install bootstrap, just have a copy of your wpa_supplicant file somewhere as it will wipe that out. You should be able to just run the install bootstrap and be good to go
katie disimone
@Kdisimone
Sep 19 2017 18:15
@Elisakelley here's my thoughts...unverified. When wifi or BT isn't behaving as it should, I check the usual files. If those all look good and it appears to be a problem with bnep0 or wlan0 service, I tend to reflash entirely. Just because I think those services are a jubilinux issue rather than an oref0 issue
But I am also quite comfortable with flashing, so I tend to have no issues going to that as a solution.
Nalla306
@Nalla306
Sep 19 2017 18:16
managed to flash cc1110......just power light on now but still same issue, constant reboot to fix radio error loop, checked the boards in place and seems ok, removed battery etc. Bugger
Elisakelley
@Elisakelley
Sep 19 2017 18:23
@Kdisimone reflash even with a pre-flashed board?
katie disimone
@Kdisimone
Sep 19 2017 18:24
Yup. I've done that plenty of times.
Nalla306
@Nalla306
Sep 19 2017 18:45
any way i can start from scratch? as i cant do anything now as soon as i connect and login it starts broadcasting rebooting to fix radio error and gives me 2 minutes
Brian Rabinovitz
@bdr1177
Sep 19 2017 18:54
What's the best way to make multiple PRs to different parts of the docs? Do I use the same fork, or do I delete the fork and re-fork every time I make another PR?
Scott Leibrand
@scottleibrand
Sep 19 2017 18:58
You'll want to make a new branch (not a new fork) for each PR.
Dana Lewis
@danamlewis
Sep 19 2017 18:58
If it's major changes - i'd create a new branch in your fork, and then another Pr for that branch into master. If it's minor stuff, you can group it together in the same Pr (and branch)
That makes it easier to discuss substantial changes in each PR and not hold one section up for another
Scott Leibrand
@scottleibrand
Sep 19 2017 18:59
@nalla306 you can do shutdown -c to cancel the reboot and service cron stop to prevent it from issuing another one while you troubleshoot.
Also, Ctrl-L will redraw the screen if a shutdown message pops up and gets in your way.
Nalla306
@Nalla306
Sep 19 2017 19:01
cheers....can i just confirm i have both usb's connected to pc and the battery connected too. It's strange as it all seemed to be showing in nightscout and its happening on both rigs. I must be doing something silly
and the wifi network i chose has to be the same as the one my laptops on?
Brian Rabinovitz
@bdr1177
Sep 19 2017 19:07
The PR instructions start with making a fork of the master docs and then create a branch by initiating an edit (in one of two ways: either from the docs or directly in github). But if I initiate an edit a second time, it's still in the same branch. How do I make the edit initiate in a new branch?
Maybe I found it. Create a new branch under the branch dropdown menu? I guess I should make a new branch for each edit so that they aren't carried across the different branches. Since I created a pull request in my master branch, I'll wait for that to be processed and then refork and start creating new branches for each edit. Is this right?
katie disimone
@Kdisimone
Sep 19 2017 19:35
If you initiate a second edit FROM THE OPENAPS DOCS DIRECTLY, it will create a new branch for the PR as soon as you save the edit.
Brian Rabinovitz
@bdr1177
Sep 19 2017 19:38
Thanks @Kdisimone
katie disimone
@Kdisimone
Sep 19 2017 19:39

I find that starting PRs from your own fork @bdr1177 is a less-than-optimal way to PR. If your master branch is out of sync with updated docs...then your PR could contain errant info that doesn't jive with current docs.

So best to start your PR by instead, (1) finding the page online in the OpenAPS docs themselves and using the little "edit in github" link in top corner. That will automatically open the file in github and (when you save the edit), create a new branch from the current up-to-date docs with your change reflected. The new branch will be named something like patch-1. Then you can do a PR from patch-1

Brian Rabinovitz
@bdr1177
Sep 19 2017 19:42
@Kdisimone Okay. I haven't PRed in a while. I don't think the directions I originally used had the instructions to PR directly from the docs. I'll try that from now on.
katie disimone
@Kdisimone
Sep 19 2017 20:21
@bdr1177 yup, old docs had that differently. New docs spell it out a little cleaner. http://openaps.readthedocs.io/en/latest/docs/Resources/my-first-pr.html
Brian Rabinovitz
@bdr1177
Sep 19 2017 20:34
@Kdisimone Yep, I read the new docs, but it still gives the option of two ways to do it. So I just did it the way I originally learned to do it.
katie disimone
@Kdisimone
Sep 19 2017 20:51
@bdr1177 actually the new docs only spell out the one method. The forking is only for the first PR. If you scroll down, the docs say that future PRs will build off a new branch in your existing for. Not sure where you are seeing the old method (starting a PR from your master branch) is in the docs. If you see that, can you link it here so we can get that fixed up? Thanks.
*your existing fork
Brian Rabinovitz
@bdr1177
Sep 19 2017 21:05
@Kdisimone hmm. I'll have to look later. I'm sure you're right and I probably just read through too fast while simultaneously relying on my memory of the old instructions.
Elisakelley
@Elisakelley
Sep 19 2017 21:21
@Kdisimone @scottleibrand @davidkeddydb thank you guys so much for your help! I re-flashed and started from scratch and everything is working well now!
garykidd
@garykidd
Sep 19 2017 21:22
@scottleibrand meant 'any' not 'and' Was jus thinking that ability to cancel what would otherwise result in an SMB based on current BG might be something to consider rather than canceling based on where BG is expected to go....seems like it might be a small price to pay in terms of responsiveness...but with a fair degree of safety added as a result. Wondering if you agree---and if I'm articulating this well enough even?
Scott Leibrand
@scottleibrand
Sep 19 2017 21:30
I would do both. The predBG array would start from the current BG, so current BG being low or any future predBG would prevent SMBs.
garykidd
@garykidd
Sep 19 2017 21:31
that seems like a significant improvement. Thanks, Scott.
One doesn't need a separate API token/key and/or user key for each rig...right?
katie disimone
@Kdisimone
Sep 19 2017 21:34
@garykidd correct
Marco
@CaptainBalou
Sep 19 2017 22:27
Hello. I am trying 0.6.0-dev a bit but I am facing errors. After drilling down that I think my problem ist this error:
root@captainsloop1:~/myopenaps# openaps monitor-pump
pump://JSON/read_clock/monitor/clock.json
reporting monitor/clock.json
pump://JSON/read_temp_basal/monitor/temp_basal.json
reporting monitor/temp_basal.json
pump://JSON/iter_pump_hours/monitor/pumphistory.json
reporting monitor/pumphistory.json
tz://JSON/rezone/monitor/pumphistory-zoned.json
reporting monitor/pumphistory-zoned.json
tz://JSON/clock/monitor/clock-zoned.json
reporting monitor/clock-zoned.json
iob://text/shell/monitor/iob.json
No duration found for 0 U/hr basalundefined { _type: 'PumpSuspend',
  _description: 'PumpSuspend 2017-09-19T18:15:08 head[2], body[0] op[0x1e]',
  timestamp: '2017-09-19T18:15:08+02:00',
  _body: '',
  _head: '1e01',
  _date: '884f121311' } { _type: 'TempBasal',
  temp: 'absolute',
  _description: 'TempBasal 2017-09-19T18:15:03 head[2], body[1] op[0x33]',
  timestamp: '2017-09-19T18:15:03+02:00',
  _body: '00',
  _head: '3300',
  rate: 0,
  _date: '834f121311' } { _type: 'BasalProfileStart',
  _description: 'BasalProfileStart 2017-09-19T18:15:03 head[2], body[3] op[0x7b]',
  timestamp: '2017-09-19T18:15:03+02:00',
  _body: '243200',
  profile_index: 18,
  _head: '7b12',
  rate: 1.25,
  offset: 64800000,
  _date: '834f121311' }
reporting monitor/iob.json
pump://JSON/reservoir/monitor/reservoir.json
reporting monitor/reservoir.json
pump://JSON/read_battery_status/monitor/battery.json
reporting monitor/battery.json
pump://JSON/status/monitor/status.json
reporting monitor/status.json
root@captainsloop1:~/myopenaps#
Can someone give me a hint to fix this - if that is an error I have to care for at all...
Scott Leibrand
@scottleibrand
Sep 19 2017 23:04
@CaptainBalou my first guess is that you changed basal profile at the exact same time you (or OpenAPS) were setting a temp basal to zero, so the temp basal record and the accompanying temp basal duration record got split apart, and it's failing to find the accompanying one.
Marco
@CaptainBalou
Sep 19 2017 23:14
Hi @scottleibrand . Nice to hear you. Thanks for digging in. No I didn't change the profile at this time. Not today at all.
But I suspended the pump at 18:15 today. Yes.
Scott Leibrand
@scottleibrand
Sep 19 2017 23:15
OK. do you know what the BasalProfileStart record at 2017-09-19T18:15:03 was then?
the PumpSuspend was at 2017-09-19T18:15:08 (5 seconds later)
Marco
@CaptainBalou
Sep 19 2017 23:15
I just canceled temp basal because someone told me to do that before suspending.
I always do that before suspending.
Scott Leibrand
@scottleibrand
Sep 19 2017 23:16
the best practice is to set a temp to zero and then suspend. whether or not you cancel before setting the temp to zero is just whichever is more convenient: doing a cancel or scrolling the temp rate down to zero
Marco
@CaptainBalou
Sep 19 2017 23:18
Yes this is what I do, but to be honest I think this was not what I did today at 18:15. I think I remember I just canceled tbr. :-(
But I think this is just a calculation which is going wrong, isn't it? It should not result in such errors, isn't it?
But after a few minutes after starting with 0.6.0-dev it seems that I am looping now even the error is still visible in the log.
Scott Leibrand
@scottleibrand
Sep 19 2017 23:20
you've apparently found a bug in our pumphistory parsing that is triggered by the particular sequence of history events you managed to generate. glad to hear it's not preventing it from looping.
I doubt this is a regression in 0.6.0-dev: the same thing would probably happen in 0.5.3
Marco
@CaptainBalou
Sep 19 2017 23:21
I first thought it prevents me from looping but it doesn't.
Scott Leibrand
@scottleibrand
Sep 19 2017 23:21
you can test that on another rig though if you like
Marco
@CaptainBalou
Sep 19 2017 23:21
I just have this one, sorry...
Never seen this before in 0.5.3
But never watched out for it for sure.
Scott Leibrand
@scottleibrand
Sep 19 2017 23:21
ok. you could set up a second install in another directory and see if it has the same error on the same pumphistory
I suspect it's a rare thing triggered by a race condition in writing the history records on the pump itself, and then decocare/openaps failing to parse it
if so, the oref0 version wouldn't matter
Marco
@CaptainBalou
Sep 19 2017 23:22
I had trouble last time switching back to 0.5.3 - had to reflash there. So at this point of time I am just happy in having it looping. ;-)
Scott Leibrand
@scottleibrand
Sep 19 2017 23:23
so if you're ok just watching to make sure it doesn't repeat, that's fine.
you might also want to work on building another rig
Marco
@CaptainBalou
Sep 19 2017 23:23
Let's see. ;-)
I have seen that part of determine_basal.js I had a modificaton in has been modified in 0.6.0.
I have to investigate how I can get my changes in there again.
But I think maxSMBBasalMinutes does something similar to what I did.
My problem with SMB is that it is much to conservative. My basal rate sometimes is 0.3 and so SMB tries to push me down from a ton of pizza with increments of 0.1 U which never works.
So I modified the maxBasal = currentbasalrate /2 to 3 which worked perfect for me. Now the code is changed here. grrrr*
:-)
Marco
@CaptainBalou
Sep 19 2017 23:33
But FIRST I have to go to bed. 01:32 am here. See you and thank again for your assistance!
Scott Leibrand
@scottleibrand
Sep 19 2017 23:33
yeah, if you want a larger maxBolus, you can increase maxSMBBasalMinutes
there is also some code to allow it to use insulinReq/2 instead of /3 when appropriate, which might help there too
specifically, when you haven't yet dosed insulin for all of your meal carbs, it does the more aggressive insulinReq/2 calculation
sleep well. :)
Martin Haeberli
@mhaeberli
Sep 19 2017 23:59
off topic - trying out Dexcom API, but getting stuck at how to even set / create a redirect-uri - that is, what value would I like to define for my new application ...