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

17th
May 2017
kenkotch
@kenkotch
May 17 2017 00:15
@scottleibrand thanks
tynbendad
@tynbendad
May 17 2017 01:08
best place to buy edison compute module?
Dana Lewis
@danamlewis
May 17 2017 01:09
Anywhere for unflashed. Hamshield for pre-flashed w Jubilinux
tynbendad
@tynbendad
May 17 2017 01:10
first one i got at amazon for $38, now they're all $53.5... seems like collusion going on :)
Dana Lewis
@danamlewis
May 17 2017 01:12
And the low power ones often go for more even though no one can tell a difference
(But yea, shop around - they're all the same from what we can tell)
Sgoldberg89
@Sgoldberg89
May 17 2017 01:14
Hi all, I am getting the following error message in my Tail log: Not an openaps environment, run: openaps init
when I run openaps init I get a list of items and can't tell if anything is amiss.
Scott Leibrand
@scottleibrand
May 17 2017 01:22
easiest solution for git errors like that is to rm -rf .git and let oref0-reset-git recreate it
Sgoldberg89
@Sgoldberg89
May 17 2017 01:24
Thanks @scottleibrand I think old faithful just went down, can't connect via ssh anymore :(
Sgoldberg89
@Sgoldberg89
May 17 2017 01:40
@bdr1177 @mhaeberli Hey, was looking through old convos and noticing that others are having issues with the radio board on the Edison. I can fix mine but it breaks again within 20-60 minutes of fixing it. Wondering if any of you are experiencing something similar.
David Rimmer
@zjedr
May 17 2017 01:40
just reviewing oref1, is there a doc yet on its features and enabling it? I particularly like the idea of UAM, but can't see what triggers a UAM.
Preferences also has 'remainingCarbsCap', any expatiation for this setting?
Will new features be able to be triggered when running offline, ie 'eating soon' via json update or from my pump, could put in 20g carb and 0.1u (as pump needs to dose to send info) to trigger smb_COB?
Sgoldberg89
@Sgoldberg89
May 17 2017 01:46
@zjedr Can you point me in the direction of the Oref1 setup info? I am interested in learning more about it but haven't been able to find it
Dana Lewis
@danamlewis
May 17 2017 01:51
@zjedr no docs yet. Check out the PRs and issues mentioned earlier today for background. Depending on which preferences you toggle, yes that'll work offline
Dana Lewis
@danamlewis
May 17 2017 01:52
(Scroll up from this afternoon for some other relevant links)
Sgoldberg89
@Sgoldberg89
May 17 2017 02:03
I keep finding the same pages that explain the reasoning behind SMB but nothing regarding how to use it/test it
I'm sure that's not an accident :)
Paul Dickens
@thebookins
May 17 2017 02:06
Any experts on bluetooth pairing out there? I'm trying to pair the G5 with Edison and I think that the pairing request doesn't have the required security level. Wondering whether a kernel upgrade will help.
Any advice appreciated - getting out of my depth!
kenkotch
@kenkotch
May 17 2017 02:25
when using an explorer board am I correct that I want radio_type subg_rfspy?
What port do I want?
Dana Lewis
@danamlewis
May 17 2017 02:25
you want to cancel out, go back again and say “yes” to explorer board
Eric
@ecc1
May 17 2017 02:26
@thebookins Not an expert, but when I worked on BLE for the G4 Receiver, I stared at a lot of hcidump, dbus-monitor, and busctl output
kenkotch
@kenkotch
May 17 2017 02:27
I did say yes to explorer board. I keep getting
"Could not get subg_rfspy state or version. Have you got the right port/device and radio_type?"
when listening
Dana Lewis
@danamlewis
May 17 2017 02:29
k. that’s an error that should self-resolve
Eric
@ecc1
May 17 2017 02:29
@thebookins You can also use wireshark to look at traffic captured by hcidump in a more friendly way
kenkotch
@kenkotch
May 17 2017 02:30
Thanks. I've waited over an hour and no looping. Starting to search for my errors
Scott Leibrand
@scottleibrand
May 17 2017 02:33
@kenkotch what are your two LEDs next to the microUSB ports doing?
they should blink when you mmtune and then go off. if they stay on solid, you need to run ccprog to re-flash your cc1110 chip on the Explorer board
Paul Dickens
@thebookins
May 17 2017 02:34
Thanks @ecc1 I'm already using wireshark (I think you suggested it) and have been staring at a lot of hex bytes. I think I've narrowed down the problem to the SMP pairing request being sent to the G5. On my mac it's 01 04 00 05 10 03 03 but on the Edison it's 01 04 00 01 10 00 01. Difference being MITM bit is set on the Mac and the Mac is offering to distribute more keys (bytes 6 and 7).
Scott Leibrand
@scottleibrand
May 17 2017 02:34
@kenkotch if that's what you're seeing, we could use some help PR'ing the troubleshooting or other appropriate sections of the docs to make that easier to find
Paul Dickens
@thebookins
May 17 2017 02:34
From my reading later versions of the kernel support higher BT security.
Sound plausible?
kenkotch
@kenkotch
May 17 2017 02:35
@scottleibrand they each blink once then go off, then I get a giant list under Traceback
Happy to help PR as soon as I get it running
I am guessing my issues are coming from the Explorer. I have 2 edisons and one explorer - same issues. New board coming end of week/next week. Do you think it is worth reflashing anyway?
Eric
@ecc1
May 17 2017 02:42
@thebookins Definitely sounds plausible. I've run into other limitations in the relatively ancient Edison kernel. But maybe it's possible to set that bit via some bluez dbus interface?
Scott Leibrand
@scottleibrand
May 17 2017 02:46
@kenkotch worth a shot. I'd try a simple reseat of the Edison on the Explorer board first though.
kenkotch
@kenkotch
May 17 2017 02:47
thanks. ill try both
kenkotch
@kenkotch
May 17 2017 03:07
reseated, reflashed. no dice. calling it quits for the night. I'll come back (hopefully) with more info tomorrow
Dana Lewis
@danamlewis
May 17 2017 03:15
@Sgoldberg89 not an accident ;)
however, would you like fries with that? ;)
Starting to rough out all of the oref1/SMB/UAM testing stuff in one page, since I know it’s a bit of a rabbit hole. This is NOT ready for merging/primetime, but hopefully a little bit easier to read on one page. Feedback welcome for what other useful things have been said in channel that should be added to this page. https://github.com/openaps/docs/blob/3421cffdbd16f77275a08da3a58cb098a947db2a/docs/docs/walkthrough/phase-4/oref1-features.md
(it’s part of a PR that won’t be merged until dev goes to master; at that point will be updated with how SMB/UAM stuff will be enabled in master; but for now it’s a cheat sheet of the dev testing instructions)
(you can also PR to that page and propose edits/additions via the max-COB-addition branch)
Martin Haeberli
@mhaeberli
May 17 2017 03:19
@Sgoldberg89 at the moment - the radio board in my case seems to be pretty stable … / I do have to reboot sometimes ; but this is more associated with the USB / serial cable / port having intermittent issues
Dana Lewis
@danamlewis
May 17 2017 03:31
@jbwittmer how’s it going?
@Kdisimone know you’re on low battery/offline, but when back you may want to eyeball & add to ^^. I know we’ve had some good back channel discussions and you may remember which bits are useful for general population in that type of page.
tynbendad
@tynbendad
May 17 2017 03:33
the insulin didn't get there from the pump?...
    "enteredBy": "openaps://medtronic/523",
    "insulin": null
Paul Dickens
@thebookins
May 17 2017 03:33
Thx @ecc1 I'll look into the dbus interface. Worth a shot.
Scott Leibrand
@scottleibrand
May 17 2017 03:35
@tynbendad I'd need more context to know what that means
tynbendad
@tynbendad
May 17 2017 03:37
sorry, that was from carbhistory.json... about 3U, the carbs got seen and sent to NS but the insulin never got to NS. i'm pretty sure the IOB from openaps is correct though...
Scott Leibrand
@scottleibrand
May 17 2017 03:38
I don't think carbhistory gets used if pumphistory has carbs in it
tynbendad
@tynbendad
May 17 2017 03:39
ok, i do see the insulin in pumphistory... but didn't get to NS
Jason Wittmer
@jbwittmer
May 17 2017 04:05
@danamlewis Thanks for asking. An unfortunate timed episode of ??? for Dexcom sensor. So. Not that well. But up until that point was going nicely. I'm more or less being a SMB surrogate with q15 minute FSBS and back of the envelope calculations until we get back online. Will keep you informed and try again tomorrow night with (hopefully) a more reliable Dexcom site.
Dana Lewis
@danamlewis
May 17 2017 04:06
womp womp. those terrible ???s :(
Jason Wittmer
@jbwittmer
May 17 2017 04:06
Couldn't have been more unfortunately timed. And because we're on pediatric receiver can't use RAW data either. Ugh.
Dana Lewis
@danamlewis
May 17 2017 04:06
double ugh :(
(for anyone who doesn’t know - we added the ability to loop off of raw data up to something like 160 mg/dl for times like this, using the g4-upload option and a plugged in g4 receiver…but has to be an adult receiver :( )
tynbendad
@tynbendad
May 17 2017 04:10
can you loop off of raw data from a g5 with xdrip+?
Scott Leibrand
@scottleibrand
May 17 2017 04:11
@jbwittmer it might be worth obtaining an adult receiver. you can use it with your current transmitter without affecting your other receivers.
Jason Wittmer
@jbwittmer
May 17 2017 04:11
But by the way @danamlewis, I know you and Scott hear this way too often, and maybe it stops even being meaningful after a while, but thanks.. From the bottom of my heart. What you do is so meaningful for us. When I first heard about oref0 I thought I could never do it and now we're nearly coming up on a year looping. When I heard about oref1 I told my wife, never. Not ever. Too risky. And yet here I am and based on the time it was working tonight really excited for where I go next. So keep it up. Hope we get to meet in person some day so you can meet my sons and we can say thanks in person. Have a good night.
@scottleibrand Already on my list of "to do's"
Scott Leibrand
@scottleibrand
May 17 2017 04:12
:-D <3
that kind of thanks is extremely meaningful. dunno if you saw Dana's blog post on that topic recently, but seeing the difference our work is making in people's lives is what keeps us doing it.
Dana Lewis
@danamlewis
May 17 2017 04:13
It does not ever stop being meaningful. Thank you for saying it <3 :smile: :sparkles:
:raised_hands: ^^
Sgoldberg89
@Sgoldberg89
May 17 2017 04:14
@danamlewis OMGTHANKYOU!
Dana Lewis
@danamlewis
May 17 2017 04:15
@tynbendad not sure if raw coming from xdrip is the same as raw off a receiver? (Dunno, actually, not sure anyone's tried it)
Scott Leibrand
@scottleibrand
May 17 2017 04:15
I recently saw @ps2 say the same thing on FB: that this is the kind of thing that gets him to keep coding. thanks to everyone who shares feedback and pays it forward however they can.
Dana Lewis
@danamlewis
May 17 2017 04:15
@Sgoldberg89 whoa - they're not that good; don't be TOO excited ;) like I said - feedback on what's missing is needed! ( if you're welcome :smile: )
Sgoldberg89
@Sgoldberg89
May 17 2017 04:16
I know, it was just a nice treat to come back to from dinner

rushofemotions

Dana Lewis
@danamlewis
May 17 2017 04:16
K :smile:
Scott Leibrand
@scottleibrand
May 17 2017 04:16
:-D #truethat
Martin Haeberli
@mhaeberli
May 17 2017 04:24
sudo bt-pan client AA:BB:CC:DD:EE:FF
Traceback (most recent call last):
  File "/usr/local/bin/bt-pan", line 228, in <module>
    if __name__ == '__main__': sys.exit(main())
  File "/usr/local/bin/bt-pan", line 196, in main
    try: iface = net.Connect(opts.uuid)
  File "/usr/lib/python2.7/dist-packages/dbus/proxies.py", line 145, in __call__
    **keywords)
  File "/usr/lib/python2.7/dist-packages/dbus/connection.py", line 651, in call_blocking
    message, timeout)
dbus.exceptions.DBusException: org.bluez.Error.Failed: Input/output error
any hints?
i keep on going back to killall bluetoothd ...
Sgoldberg89
@Sgoldberg89
May 17 2017 04:32
Am I supposed to see options re enableUAM in preferences.json or add them manually? I am wondering if I didn't instal the dev version as I though I did or if I am looking in the wrong preferences.
Dana Lewis
@danamlewis
May 17 2017 04:33
Did you npm run global install
(You should see them; so you probably skipped somewhere)
Sgoldberg89
@Sgoldberg89
May 17 2017 04:34
thought I installed the dev version
Scott Leibrand
@scottleibrand
May 17 2017 04:38
Did you re-run oref0-setup? That is what (re)creates preferences.json.
Sgoldberg89
@Sgoldberg89
May 17 2017 04:38
Is it ok to have multiple loops with 1 running Oref1 and another running Oref0?
Scott Leibrand
@scottleibrand
May 17 2017 04:38
You should pause one of them.
Sgoldberg89
@Sgoldberg89
May 17 2017 04:38
I may have deleted them. I am starting over to make sure
Ok.
Tomorrow I'm going to ask about what exactly I need to do in nightscout/heroku. But I'm going to spread it out over two days so it seems less egregious to the keepers of the keys
Scott Leibrand
@scottleibrand
May 17 2017 04:44
Lol
Someone who keeps the NS keys (besides me, since I submitted it) should just go merge that PR. And to encourage them, everyone who's been running the floating-carbs branch of NS without issue should go :+1: it.
Martin Haeberli
@mhaeberli
May 17 2017 04:50
help w / bluetooth?
still same problem as above
Scott Leibrand
@scottleibrand
May 17 2017 05:00
Nobody here really seems to understand Bluetooth. We got a process that works by trial and error, but that didn't teach us much for trying anything new.
Martin Haeberli
@mhaeberli
May 17 2017 05:01
@scottleibrand thx - fwiw - i tried following the process VERY carefully 10+ times (showing i’m insane maybe?) always the same outcome
ok, back at “bg data too old” despite relatively recent bg data available through ns api
and my usual run openaps ns-loop by hand trick not working
Paul Dickens
@thebookins
May 17 2017 05:06
@mhaeberli Are you running xdripAPS? If so, and you're not looping locally, the new ns data can get clobbered by the old xdrip data, leading to the "bg data too old" message.
Just a thought.
Need to modify the cron job to compare data recency.
Martin Haeberli
@mhaeberli
May 17 2017 05:07
@thebookins thanks - don’t think so - g4 -> receiver -> iphone -> dexcom share cloud -> nightscout -> edison openaps
Paul Dickens
@thebookins
May 17 2017 05:07
got it
no idea sorry, we're on G5
Scott Leibrand
@scottleibrand
May 17 2017 05:10
@thebookins would you be able to help get oref0-setup updated to fix that for xdrip users if it's not already fixed in dev?
Paul Dickens
@thebookins
May 17 2017 05:12
sure @scottleibrand I know we talked about that a few weeks ago, but haven't quite got there yet. Will do.
Scott Leibrand
@scottleibrand
May 17 2017 05:19
:+1:
Dana Lewis
@danamlewis
May 17 2017 05:21
Thank you @thebookins for everything you're doing!
Paul Dickens
@thebookins
May 17 2017 05:23
:+1: thank you too @danamlewis and @scottleibrand , awesome community to be part of, with real tangible benefits for our son. Couldn't imagine not looping now!
Scott Leibrand
@scottleibrand
May 17 2017 05:24
:smile:
Dana Lewis
@danamlewis
May 17 2017 05:24
:heart: me either :smile:
Tim Street
@tim2000s
May 17 2017 05:26
@scottleibrand @thebookins - that cron job I stuck in a week back kicks that one out (at least as a stop gap) so anyone running latest dev shouldn't be seeing the xdripaps caused "BG too old" issue.
Paul Dickens
@thebookins
May 17 2017 05:27
Tops @tim2000s
Martin Haeberli
@mhaeberli
May 17 2017 05:30
turns out the debugging trick at this end for bluetooth was to power cycle the iPhone. go figure
PieterGit
@PieterGit
May 17 2017 06:02
@ecc1 @thebookins https://github.com/andy-shev/linux/tree/eds kernel now has bluetooth support as well. didn't give it a try yet. Don't know if it's merged in mailine v4.12-rc1 yet.
Martin Haeberli
@mhaeberli
May 17 2017 06:08
I’ll try to add a PR
a PR re bluetooth debugging specs ...
Paul Dickens
@thebookins
May 17 2017 06:16
@PieterGit I know next to nothing about Linux kernels. How complicated is it to build and deploy this to the Edison? Certainly looks to me a more up to date SMP stack.
PieterGit
@PieterGit
May 17 2017 07:02
See https://communities.intel.com/thread/75472?start=60&tstart=0 I tried a while back because I thought it would help the usb otg problems. Those turned out to be hardware related with first explorer rev. If it fails you might need to do xfstk recovery.
I made some initial patches for 4.x kernels that are in dev and it should be possible to use ccxxxx chip as normal user without mraa . Ecc1 also ran some 4.x kernels
Needless to say... only try it on a spare explorer or Edison
Jason Wittmer
@jbwittmer
May 17 2017 07:05
@danamlewis After 4 hours of ??? and q15' FSBS I can fully confirm my back of the envelope calculations are crap compared to oref1 . Amazing to see it pick up once we have Cgm reading and start correcting my mistakes. Guess I'll be up all night to watch it. Reminiscent of 1 year ago when I stayed up all night and watched the first oref0 loop. Or even worse the "bad old days" before we had anything to help us and it was just wake up ever 2 hours after Chinese take out for more insulin and the inevitable overcorrection low about 4am . Maybe we'll try the same meal tomorrow night with a better functioning CGM site and see how much better it can do. In fact I think that will be a great field test. Same ill-advised meal, same up front bolus and let's compare the results.
PieterGit
@PieterGit
May 17 2017 07:20
@thebookins Looks like 4.12 will be a good future kernel for openaps but there will be quite some work to get there http://www.phoronix.com/scan.php?page=news_item&px=Intel-Edison-BT-Linux-4.12
fidoman
@efidoman
May 17 2017 11:13
image.png
Another great SMB experience with last night's meal that showed delayed absorption of carbs. SMB picked it up until all carbs were decayed while I was sleeping the meal off. Wonderful!
fidoman
@efidoman
May 17 2017 11:29
My glucose leveled off around 1am at 130 and I woke up at 110 this morning at 6am
elodaille01
@elodaille01
May 17 2017 11:35
blob
@brandonschmidt :
(about my laptop)
elodaille01
@elodaille01
May 17 2017 12:47
blob
I'm afraid Autotune is not for me :'(
MelissaParkey
@MelissaParkey
May 17 2017 13:25
@Kdisimone Back on track.... but I can't remember how to find the link to the previous Gitter stream you sent me. :worried:
Tim Street
@tim2000s
May 17 2017 13:35
@elodaille01 You could always run it up on a VM but you may have to ask in the Autotune channel for some help with that.
elodaille01
@elodaille01
May 17 2017 13:42
oooh...ok !!! Sorry !!!!!!
tynbendad
@tynbendad
May 17 2017 14:14
@elodaille01 use google cloud, it wasn't too hard. some pc's just can't do VM
@scottleibrand it happened again - carbs got into nightscout, but not insulin... is this a problem with the way he's using the pump, or an internet connection drop? IOB/COB are fine, but NS not updated, so i think that means autotune won't see the insulin... (when i see it i can fix it manually, but i don't look at all bolus').
Tim Street
@tim2000s
May 17 2017 14:18
@tynbendad have you checked whether it's in the mongo db?
tynbendad
@tynbendad
May 17 2017 14:19
checking...
katie disimone
@Kdisimone
May 17 2017 14:19
@tynbendad I've found when NS isn't displaying a bolus, it's usually still in OpenAPS. It will have the appropriate iob in the OpenAPS pill. So somehow ns has the problem in the render
tynbendad
@tynbendad
May 17 2017 14:20
it is not in mlab either
yes, its ok in openaps, but didn't make it to NS... not just a display issue, the insulin treatment didn't make it there at all. but the carbs did.
iob/cob fine
estimate is in with the carbs also "notes": "Food estimate 5\nCorrection estimate 0\nBolus estimate 5\nTarget low 80\nTarget high 88\nHypothetical glucose delta 0",
Eric
@ecc1
May 17 2017 14:58

@PieterGit @thebookins It's not hard to build just an Edison kernel (not a full flashable image) to drop into an Edison's /boot directory, and then boot it by hand in u-boot. I've done that to try patches on both 3.10 and 4.x kernels, and can help you through it.

My problem with Andy's 4.11 kernel has been that it boots OK when the Edison is on an Intel mini-breakout, but not when it's on a Sparkfun base board, for some reason that is beyond me. No idea whether it will boot on an Explorer board, but that would be an interesting data point.

Jason Wittmer
@jbwittmer
May 17 2017 15:04
Screen Shot 2017-05-17 at 10.03.17 AM.png
So that's our SMB run for breakfast this morning. Look alright to everyone? I'm still a little weirded out by the long zero basal up front but it seemed to work great.
MelissaParkey
@MelissaParkey
May 17 2017 15:20
@scottleibrand @Kdisimone !!!! ssh root@10.0.0.17 worked. entered pw and it took me to my edison!!!!!
katie disimone
@Kdisimone
May 17 2017 15:25
👍🏻
brandonschmidt
@brandonschmidt
May 17 2017 15:26
@elodaille01, your processor doesn't support virtualization natively: http://ark.intel.com/m/products/50176/Intel-Pentium-Processor-P6200-3M-Cache-2_13-GHz
This means you would have to rely upon the OS to do it for you, which is what the Hyper-V feature does. But your version of Windows
Scott Leibrand
@scottleibrand
May 17 2017 15:29
@jbwittmer there is an update coming shortly (from the pushover branch) that will help it not zero temp as aggressively up front
brandonschmidt
@brandonschmidt
May 17 2017 15:31
...doesn't support that. Even if you upgraded to Win 10 Pro, your CPU wouldn't support the Hyper-V feature because it doesn't have SLAT.
So, when you run VirtualBox, you're relying upon it's own virtualization implementation, as oppsed
...to Windows' implementation or the CPU's.
And VirtualBox is only going to support 32 bit in those circumstances.
So, you should reach out to the Autotune people for alternatives to running VirtualBox on your local computer.
Scott Leibrand
@scottleibrand
May 17 2017 15:36
It's probably far simpler to just run a cloud VM as per the docs
brandonschmidt
@brandonschmidt
May 17 2017 15:37
Yep, in @elodaille01 case, that's the only option (or use a different hardware).
Sgoldberg89
@Sgoldberg89
May 17 2017 15:47
Good morning @scottleibrand, I was looking for the PR you made re nightscout SMB instructions and couldn't find it. Would you be so kind as to let me know where one might find them?
Scott Leibrand
@scottleibrand
May 17 2017 15:50
that one is linked from Dana's new write-up. look for cgm-remote-monitor
Dana Lewis
@danamlewis
May 17 2017 16:03
@elodaille01 you can run Autotune on your rig. Don't need to be running t on the laptop or a VM. Phase A instructions is what you want to follow.
Sgoldberg89
@Sgoldberg89
May 17 2017 16:06
So... What's an "enable flag?"
Martin Haeberli
@mhaeberli
May 17 2017 16:13
so - for my problems with stale BG, sometimes invoking
openaps bg-fresh-check
openaps get-ns-bg
Jason Wittmer
@jbwittmer
May 17 2017 16:14
Enable flag is used in oref0-setup.sh to enable advanced features
Martin Haeberli
@mhaeberli
May 17 2017 16:14
fixes it - but I still have to look for root cause; openaps ns-loop was also maybe involved in ‘fixing'
Sgoldberg89
@Sgoldberg89
May 17 2017 16:14
@jbwittmer Is that just when you select "yes" for advanced features?
during the setup process?
Jason Wittmer
@jbwittmer
May 17 2017 16:18
the enable flag will be the end result of the questions that you answer during the setup (if you're using the step by step process). Look at the contents of your oref0-runagain.sh and you can see where -enable is used
Sgoldberg89
@Sgoldberg89
May 17 2017 16:20
@jbwittmer thanks
Martin Haeberli
@mhaeberli
May 17 2017 17:14
to re-iterate a previous comment - in addition to a parameter like max_iob, it would be nice to have a parameter like min_basal_rate, can default to zero, but maybe can be set to, say 0.1 U / hr if user wants. yes, I know I can code this myself… discuss? (concerned about ‘washout’ after long zero-temp)
Dana Lewis
@danamlewis
May 17 2017 17:14
Describe what you mean by washout?
Martin Haeberli
@mhaeberli
May 17 2017 17:15
@danamlewis @bewest described washout to me last year as , say you have turned basal off for a long time 3 - 6 hours - no more insulin there, so no chance to bring even trace glucose into cells...
that was my minimal understanding at the time
I love @bewest and / but I gather he has been busy elsewhere...
i may have the description wrong
Dana Lewis
@danamlewis
May 17 2017 17:18
Yea. I don't think that's an increased risk a OpenAPS. If it was, a min 0.1 temp wouldn't do much to resolve that regardless. I think that would hamper OpenAPS ability to ward off lows for you instead
Martin Haeberli
@mhaeberli
May 17 2017 17:18
ok thx
Dana Lewis
@danamlewis
May 17 2017 17:18
What you might want to do instead is think about what set of alerts do you want
Martin Haeberli
@mhaeberli
May 17 2017 17:18
:+1:
Dana Lewis
@danamlewis
May 17 2017 17:18
I.e. Have NS alert if >1 hour low, or >2 hours low, etc
You can then manually decide if you want to trade off low temps for carbs, or set temp targets, etc.
Martin Haeberli
@mhaeberli
May 17 2017 17:19
here at the moment we don’t have a low but we have a predicted low - so bg about 107 but prediction for low so basal zero temp
but i see what you mean
Dana Lewis
@danamlewis
May 17 2017 17:20
Right - but it's not going to zero temp you for 3 hours
Martin Haeberli
@mhaeberli
May 17 2017 17:20
corrolary - low BG suspend overnight helps sleep; sustained during the day - better response is carbs, keep some insulin going
Dana Lewis
@danamlewis
May 17 2017 17:21
Even an hour or two of low temps isn't going to cause a huge problem - it's safer to deal w a bit of a rebound hours after that, vs limit your ability to save self from a low, which is the bigger short term problem
Martin Haeberli
@mhaeberli
May 17 2017 17:21
to paraphrase steve jobs, “we could be holding it wrong” - so far, zero temp for an hour or so...
k
@danamlewis thx 4 all you do - seriously ! appreciate the guidance
Dana Lewis
@danamlewis
May 17 2017 17:21
But chances are you're not going to be zero'd for hours, and if you are, you may want to think about manual actions and also evaluating underlying basals, isf, dia, etc.
:+1:
Martin Haeberli
@mhaeberli
May 17 2017 17:22
cn u pls also remind us where to register as successfully looping 3 nights (on openaps) - I’ll try to get our T1 to do this with me tomorrow
(yes, in red)
RTD
Dana Lewis
@danamlewis
May 17 2017 17:23
Haha. It's a page in phase 4 sections of docs that says "now tell us you're looping!"
Martin Haeberli
@mhaeberli
May 17 2017 17:23
good enough hint!
jdunn42
@jdunn42
May 17 2017 18:20
My kid's phone has a bad habit of failing to actually share its data connection via the bluetooth tether. Because of this, I'm going to switch to using xdripAPS. Is there any reason that I can't add out NS server URL to the REST URL field after those of her OpenAPS rigs to have XDrip continue to upload to NS?
Tim Street
@tim2000s
May 17 2017 18:27
@jdunn42 If you do that and one of the rest destinations can't be reached, I don't think it writes to either. You might want to do a mongo direct upload to NS to avoid that issue.
jdunn42
@jdunn42
May 17 2017 18:29
So it's all or nothing? Like if she only has one of her two rigs with her, it won't send the data to it? A mongo upload is fine for the NS part, but she has two Edison rigs and I don;t want to make her have to carry both all the time.
Tim Street
@tim2000s
May 17 2017 18:32
Yes. It only works on one.
Well, it only works on one at a time that should be.
PieterGit
@PieterGit
May 17 2017 18:45
@ecc1 @thebookins I had success on an explorer board and kernel 4.x and even had succesfull USB and TI communication (openaps use pump model) , but there are GPIO stuff things that give quite a lot of errors in the logs and probably need to be rewritten for use with a 4.x kernel. It's on my wishlist, but not very high, because I first want token based authentication with Nightscout fixed.
kenkotch
@kenkotch
May 17 2017 19:05

hey all- I'm finally getting somewhere (I think) in troubleshooting. In ns-loop i keep getting:
May 17 12:55:03 pinky ns-loop.log: Starting ns-loop at Wed May 17 12:55:03 MDT 2017:
May 17 12:55:19 pinky ns-loop.log: ns://JSON/shell/cgm/ns-glucose.json
May 17 12:55:19 pinky ns-loop.log: No JSON object could be decoded

ns-glucose.json has no content. glucose.json is mostly up to date. Right now it is 17min behind. Not sure where to look next. Any ideas?

Background: on my NS site I do not see the temp basals displayed, just the normal and the pump pill is all red letters and missing how long ago it communicated, but accurately shows insulin left

katie disimone
@Kdisimone
May 17 2017 19:15
@kenkotch if you message me your URL for NS, I can take a look at your settings and see if you're missing something
kenkotch
@kenkotch
May 17 2017 19:15
@Kdisimone thanks
Lynne Butts
@flutelynne_twitter
May 17 2017 20:47
I have been testing SMB on my Intel/Explorer board today (Yesterday evening wasn’t so grest, but today it was really working well with different scenarios.). I started out this morning, and got about seven and a half hours before I got the time off warning. I am done testing this for today, but here is my question: I use the Loop app in “open loop” mode while using OpenAPS a lot of the time. If I use oref0—set-device-clocks to reset my OpenAPS rig, will that mess up my connection with my Loop app and pump? I was hoping that by resetting the clock through Loop, that would be an easy fix, but I am not sure if that is a hindrance instead. Also, it is weird to see a 15 min or more gap in the NS pill while the pump is zero temping. It does pick up again, but is that a normal occurence?
Michael Spradling
@M1Sports20
May 17 2017 20:56
Sorry, new to this chat. Are there any ongoing projects in dev for new drivers for pumps. Such as reverse engineering the xslim x2 since it will have bluetooth
susieanneh
@susieanneh
May 17 2017 20:57
Hi everyone! To loop offline do I really just need to plug my rig to my cgm? I just received my hardware this month. Work has terrible wifi and I'm trying to see what would work best.
John Seaber
@jseaber
May 17 2017 21:00
@M1 @M1Sports20 : Yes, check https://omniaps.slack.com
Michael Spradling
@M1Sports20
May 17 2017 21:00
jstevensog: Thanks
Dana Lewis
@danamlewis
May 17 2017 21:01
@m1sports20 do join the above channel; but main ongoing work is around Omnipod. We don't have any reason to believe we'd be able to remotely set temps on the x2 or other t:slims
@susieanneh add a second battery too
susieanneh
@susieanneh
May 17 2017 21:01
Awesome 🙏🏻 thanks
Dana Lewis
@danamlewis
May 17 2017 21:02
@flutelynne_twitter not sure if setting clock via Loop will mess it up but worth a try. (We have an automated fix for that coming soon, too, but you can try that in meantime). Also which pill are you talking about being delayed?
Lynne Butts
@flutelynne_twitter
May 17 2017 21:07
The OpenAPS pill will sometimes show a delay of 12-16 minutes during a zero temp basal. So then no view of anything on NS. But maybe that was the start of my time being off. Yesterday evening, it was “delayed” by 2 hours, and my BGs went above 200. It was microbolusing before that happened, though. I tried the time reset as posted above, but received “null”. So I shut down the rig, and just did the Loop app.
IMG_1762.PNG
I guess that it was both pump and openaps pills. Pic is not showing up
Dana Lewis
@danamlewis
May 17 2017 21:14
Was the rig online during that time?
Lynne Butts
@flutelynne_twitter
May 17 2017 21:21
Yes. Today too.
IMG_1762.PNG
^^That was last night.
IMG_1763.PNG
^^Today
Lynne Butts
@flutelynne_twitter
May 17 2017 21:27
I also saved the info from the openaps pill on my 200+ BG pic, but don’t want to bombard the page here.
susieanneh
@susieanneh
May 17 2017 21:31
any ideas of why my dexcom isnt uploading for 15 min at a time but my open aps is... my dexcom is plugged into my rig
elodaille01
@elodaille01
May 17 2017 21:37
@tynbendad @brandonschmidt @scottleibrand @danamlewis thank you
Tim Street
@tim2000s
May 17 2017 21:46
@flutelynne_twitter did you 'killall oref0-pump-loop' before trying to reset the pump clock? That's always been necessary in the past.
hilarykoch
@hilarykoch
May 17 2017 21:57
I’m trying to install dev version and running into a snag. Not sure what to do next...
Screen Shot 2017-05-17 at 5.56.30 PM.png
Lynne Butts
@flutelynne_twitter
May 17 2017 21:59
@tim2000s No, I did not. It is now in my notes to do that first. Thank you!
Tim Street
@tim2000s
May 17 2017 22:05
@flutelynne_twitter I think the script that @scottleibrand has been working on resolves that so watch this space.
@hilarykoch What does 'df -h' show you?
Lynne Butts
@flutelynne_twitter
May 17 2017 22:06
Thank you. I read this daily.
hilarykoch
@hilarykoch
May 17 2017 22:06
Screen Shot 2017-05-17 at 6.06.30 PM.png
@tim2000s Does that help?
@tim2000s Going to try something else...
Tim Street
@tim2000s
May 17 2017 22:13
I think your /tmp file needs clearing out @hilarykoch
hilarykoch
@hilarykoch
May 17 2017 22:13
yup. clearing out everything and starting fresh
Tim Street
@tim2000s
May 17 2017 22:14
If you run 'du -h | grep git' what do you get?
Brian Rabinovitz
@bdr1177
May 17 2017 22:15
If I give insulin by shot and put that in nightscout will openaps take that insulin into account?
Dana Lewis
@danamlewis
May 17 2017 22:18
@bdr1177 no
Martin Haeberli
@mhaeberli
May 17 2017 22:19
so i paired two bluetooth iphones with it, but tried to disconnect from the “old” one; it seems still to remember to try that one. where do i tell it to forget?
Dana Lewis
@danamlewis
May 17 2017 22:19
OpenAPS won't know about that insulin. You can unplug pump and give a bolus into the air to simulate the injection though
Brian Rabinovitz
@bdr1177
May 17 2017 22:22
@danamlewis Ok. Thanks.
Tim Street
@tim2000s
May 17 2017 22:24
@mhaeberli You'll need to remove the MAC address from the cron for oref0-online.
Martin Haeberli
@mhaeberli
May 17 2017 22:24
thx
Tim Street
@tim2000s
May 17 2017 22:24
Or rather change it for the other one.
Martin Haeberli
@mhaeberli
May 17 2017 22:25
ok thx
Martin Haeberli
@mhaeberli
May 17 2017 22:31

@tim2000s so I see the line

* * * * * ps aux | grep -v grep | grep -q "oref0-online AA:BB:CC:DD:EE:FF" || oref0-online AA:BB:CC:DD:EE:FF >> /var/log/openaps/network.log

but not clear on how changing this line will do the trick ...

ok maybe getting that oref0-online is the key
Scott Leibrand
@scottleibrand
May 17 2017 22:41
@flutelynne_twitter if you're having pump clock problems, you can test the newest branch with the automatic oref0-set-device-clocks fix. Just checkout and pull the pushover branch. You don't need to use any of the pushover stuff, but the clock fix is in there with it, along with some other stuff. See the PR for details.
Dana Lewis
@danamlewis
May 17 2017 22:42
(@flutelynne_twitter let us know if you do that and it works well; that will help us merge it into dev faster)
Lynne Butts
@flutelynne_twitter
May 17 2017 22:50
Thanks. I am not sure how to do that, and don’t want to mess up what I do have. I had better wait until it gets into dev.
Scott Leibrand
@scottleibrand
May 17 2017 22:56
It's no different from checking out dev. You just replace dev with pushover.
Lynne Butts
@flutelynne_twitter
May 17 2017 22:56
Thank you.
susieanneh
@susieanneh
May 17 2017 23:02
When looping offline... nothing will show in my NS correct?
Scott Leibrand
@scottleibrand
May 17 2017 23:07
Only if you have Internet (or telepathy). ;-)
susieanneh
@susieanneh
May 17 2017 23:08
Ok nope no wifi :( lol
Scott Leibrand
@scottleibrand
May 17 2017 23:08
"looping offline" covers a number of different configurations, either with or without Internet.