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

9th
May 2017
Scott Leibrand
@scottleibrand
May 09 2017 00:17
@ericfeibelman I don't think your issue is specific to SMB. the key part of your logs is this line: BG data is too old (it's probably this), or clock set incorrectly. The last BG data was read at Sun May 07 2017 07:47:00 GMT-0400 (EDT) but your system time currently is Mon May 08 2017 20:00:43 GMT-0400 (EDT)
you'll need to figure out why monitor/glucose.json has BG data from 36 hours ago.
fidoman
@efidoman
May 09 2017 00:19
when I don't use microbolus it all works fine
I've gone back-n-forth several times and each time the same date error with SMB and all works great without it (using runagain.sh).
Scott Leibrand
@scottleibrand
May 09 2017 00:20
are you using dev branch of oref0 in both cases?
fidoman
@efidoman
May 09 2017 00:30
Yes I am using the dev branch in both cases
I can reproduce it at will. 1) enable microbolus then runagain, it breaks 2) take microbolus out of runagain, it works ... every time nothing else changing.
Scott Leibrand
@scottleibrand
May 09 2017 00:33
ok. I haven't tested SMB with Enlite, so it might be something to do with that...
could be that this code doesn't handle the Enlite's date format correctly:
    # set mtime of monitor/glucose.json to the time of its most recent glucose value
    touch -d "$(date -R -d @$(jq .[0].date/1000 monitor/glucose.json))" monitor/glucose.json
could you paste one of the records from your glucose.json to compare it?
it's expecting a date field that looks like: "date": 1494259247000,
fidoman
@efidoman
May 09 2017 00:43
Sure - I have one rig broken and one working. Let me login to the broken one now and get that
{
"trend_arrow": "45_UP",
"glucose": 92,
"direction": "FortyFiveUp",
"_id": "59111100a802700f1bc96cae",
"name": "GlucoseSensorData",
"date_type": "prevTimestamp",
"_tell": 16,
"display_time": "2017-05-08T20:37:00-04:00",
"packet_size": 0,
"dateString": "2017-05-08T20:37:00-04:00",
"device": "openaps://medtronic/pump/cgm",
"date": 1494290220000,
"sgv": 92,
"type": "sgv",
"op": 46
},
Scott Leibrand
@scottleibrand
May 09 2017 00:50
ok, that looks and works fine, so that's not the problem...
fidoman
@efidoman
May 09 2017 00:53
Thank you Scott for working on this one. I'm headed to bed now, but let me know what other logs I can provide from either the broke rig and/or the working one and I'll get it for you in the morning.
Scott Leibrand
@scottleibrand
May 09 2017 00:54
I think I found it. oref0-setup overwrites the pump-loop alias for MDT pumps, to be:
"command": "! bash -c \"sleep $[ ( $RANDOM / 2048 ) ]s; until(echo Starting pump-loop at $(date): && openaps wait-for-silence && openaps get-bg && openaps refresh-old-pumphistory && openaps refresh-old-pumphistory-24h && openaps refresh-old-profile && openaps refresh-temp-and-enact && openaps refresh-pumphistory-and-enact && openaps refresh-profile && openaps refresh-pumphistory-24h && echo Completed pump-loop at $(date) && echo); do echo Error, retrying && [[ $RANDOM > 25000 ]] && ( openaps wait-for-long-silence ; openaps mmtune; sleep 5 ); done\""
vs. "command": "! bash -c \"sleep $[ ( $RANDOM / 2048 ) ]s; until(echo Starting pump-loop at $(date): && openaps wait-for-silence && openaps refresh-old-pumphistory && openaps refresh-old-pumphistory-24h && openaps refresh-old-profile && openaps refresh-temp-and-enact && openaps refresh-pumphistory-and-enact && openaps refresh-profile && openaps refresh-pumphistory-24h && echo Completed pump-loop at $(date) && echo); do echo Error, retrying && [[ $RANDOM > 25000 ]] && openaps wait-for-long-silence && openaps mmtune; sleep 5; done\""
the difference being openaps get-bg has to be run inside pump-loop with MDT CGM, vs. in its own loop for Dexcom
so we'll need to add something similar to oref0-pump-loop.sh to support Enlite
Scott Leibrand
@scottleibrand
May 09 2017 01:13
@ericfeibelman I just coded up a fix for you. Can you please test openaps/oref0#462 tomorrow?
Martin Haeberli
@mhaeberli
May 09 2017 01:44
@scottleibrand any chance that this also explains why my system fails with “old” bg data?
if so, I can try to test the fix tonight
not running MDT cgm, though
Scott Leibrand
@scottleibrand
May 09 2017 01:46
Does your setup (xdrip?) modify the pump-loop alias?
Might need a similar fix for that if so.
Martin Haeberli
@mhaeberli
May 09 2017 01:46
not as far as i know - G4 share -> dexcom share -> Nightscout ...
Scott Leibrand
@scottleibrand
May 09 2017 01:47
What did you tell oref0-setup for CGM?
Martin Haeberli
@mhaeberli
May 09 2017 01:47
but frequently / intermittently goes to a situation where Nightscout BG is current; NS Api curl fetch gets current, but cgm/glucose.json and ns-glucose.json somehow not updated
G4
ok - one of the edisons just cured itself...
let’s see about the other ...
Martin Haeberli
@mhaeberli
May 09 2017 01:52
the other is still stuck...
cgm# ls -l
total 56
-rw-r--r-- 1 root root 53623 May  8 17:10 glucose.json
-rw-r--r-- 1 root root     0 May  8 18:51 ns-glucose.json
but last NS update is 18:52 or so - so trying to figure out why glucose.json etc not updated. in this condition, my experience is that openaps ns-loop causes stuff to update ONCE, enough to get openaps happy temporarily… but of course the BG gets old again
Martin Haeberli
@mhaeberli
May 09 2017 02:11
so time stamp on ns-glucose is updated, but on glucose.json is unchanged

happy:

Starting ns-loop at Mon May 8 19:12:03 PDT 2017:
ns://JSON/shell/cgm/ns-glucose.json
No JSON object could be decoded

Refreshed temptargets and meal-carbs
{"batteryVoltage":3700, "battery":52}
Uploaded

unhappy

Starting ns-loop at Mon May 8 19:12:03 PDT 2017:
ns://JSON/shell/cgm/ns-glucose.json
No JSON object could be decoded                                                  
Refreshed temptargets and meal-carbs
{"batteryVoltage":3870, "battery":69}                                          
 Uploaded

so not seeing a difference there

(tail ns-loop.log)
Martin Haeberli
@mhaeberli
May 09 2017 02:19
so openaps get-ns-bg seems to have ‘fixed’ this on the latter instance (my prediction is, just temporarily)
Scott Leibrand
@scottleibrand
May 09 2017 02:36
You might need to remove some 2>/dev/null stuff from your aliases to see what it's really doing.
Martin Haeberli
@mhaeberli
May 09 2017 03:21
We have a loopable 522 that just started to buzz when the quick bolus button is pressed. Just ran self diagnostic - any thoughts?
@scottleibrand thx
Martin Haeberli
@mhaeberli
May 09 2017 03:50
looks like alert type got reset by self diagnostic
tepidjuice
@tepidjuice
May 09 2017 03:52
I have an openaps loopable pump which has sticky buttons i.e. the act and bolus buttons are very difficult to press and keep pressed down anyone have any experience with fixing these up?
Martin Haeberli
@mhaeberli
May 09 2017 03:54
@tepidjuice there is a link - let me see if I or somebody else can find it...
katie disimone
@Kdisimone
May 09 2017 03:54
So UAM’s likely behavoir/influence when enabled BUT meal is announced, preboluses, and carb counted? What would be different with UAM enabled in that situation.
@tepidjuice yup...http://imgur.com/a/iOXAP
I fixed mine and it was a 5-10 min easy fix
katie disimone
@Kdisimone
May 09 2017 03:57
Anyone ever see interrupting messages when logged in using screen mode? These messages come rolling by as a display every once in awhile.
root@edisonSMB:~/src/oref0/bin# cd myopenaps[58024.639148] link down if wlan0 may call cfg80211_disconnected. event : 16, reason=2 from 80:d2:1d:8c:f0:58
[58024.667721] wl_android_wifi_off in
[58024.674872] wifi_platform_set_power = 0
[58032.190539] 
[58032.190539] Dongle Host Driver, version 1.141.59 (r)
[58032.190539] Compiled in /home/robin/src/edison-linux-helper/edison-bcm43340 on Apr  4 2017 at 22:42:52
[58032.190720] wl_android_wifi_on in
[58032.211513] wifi_platform_set_power = 1
[58032.834654] F1 signature OK, socitype:0x1 chip:0xa94c rev:0x2 pkg:0x0
[58032.843180] DHD: dongle ram size is set to 524288(orig 524288) at 0x0
[58032.852413] dhdsdio_download_firmware: firmware path=/lib/firmware/fw_bcmdhd.bin, nvram path=/lib/firmware/bcmdhd.cal
[58033.029747] sdioh_request_buffer: [1] doing memory copy buf=f5265000, len=2037
[58033.041428] dhdsdio_write_vars: Download, Upload and compare of NVRAM succeeded.
[58033.212686] dhd_bus_init: enable 0x06, ready 0x06 (waited 0us)
[58033.224755] wifi_platform_get_mac_addr
[58033.224843] wifi_get_mac_addr_intel: unable to open /config/wifi/mac.txt
[58033.241045] Firmware up: op_mode=0x0015, MAC=90:b6:86:09:78:be
[58033.268983] Firmware version = wl0: Aug 15 2014 15:09:17 version 6.20.190.2 (r497061) FWID 01-bb9bbd07
[58033.279287] dhd_preinit_ioctls wl ampdu_hostreorder failed -23
[58038.378674] wl_bss_connect_done succeeded with 80:d2:1d:8c:f0:58
[58038.426464] wl_bss_connect_done succeeded with 80:d2:1d:8c:f0:58
tepidjuice
@tepidjuice
May 09 2017 03:58
@Kdisimone I thought that was mainly for moisture related issues. Does it work for generally misbehaving buttons as well?
Scott Leibrand
@scottleibrand
May 09 2017 04:00
@Kdisimone UAM predBGs will be considered alongside the COBpredBGs in that situation. It will hopefully make it less likely to overbolus if you overcount carbs, and quicker to SMB when it sees absorption from undercounted ones.
katie disimone
@Kdisimone
May 09 2017 04:00
@tepidjuice yup…dust and gunk affect it too
Scott Leibrand
@scottleibrand
May 09 2017 04:01
It might also be slightly less aggressive with UAM enabled while absorption ramps up right after eating.
tepidjuice
@tepidjuice
May 09 2017 04:01
@Kdisimone Great I'll give that a go. I want to give this one to another diabetic who's keen to try openaps.
Dana Lewis
@danamlewis
May 09 2017 04:05
@tepidjuice :clap:
katie disimone
@Kdisimone
May 09 2017 04:10
@scottleibrand thanks!
And I’m still wrapping my head around when that 120 min suspension ends. Is it when COB goes to zero or 120 min and/or something else?
Scott Leibrand
@scottleibrand
May 09 2017 04:16
It's when IOB drops low enough that IOB*ISF wouldn't drop you low enough to need more than 15m of zero temping, and eventualBG is above target.
At that point COB is usually less than 10g (or it would need to SMB more, keeping IOB up and naive_eventualBG low).
katie disimone
@Kdisimone
May 09 2017 04:17
@scottleibrand THANK you…that helps a lot
and, i’m assuming the math is subtracting out of iob the basals that it suspends as that goes along.
Martin Haeberli
@mhaeberli
May 09 2017 04:18
@Kdisimone as to ‘interrupting messages’ - yes ...
for example, when stuff is happening w/BlueTooth or WiFi
Scott Leibrand
@scottleibrand
May 09 2017 04:18
It's more complicated than that of course, but that is the gist. Yes, zero temps are included in IOB calcs.
SMBs are included in basalIOB, BTW. Manual boluses bigger than 30m worth of basal are not.
katie disimone
@Kdisimone
May 09 2017 04:27
that makes sense to me
Scott Leibrand
@scottleibrand
May 09 2017 04:50
There are also lots of cases where it will cancel the zero temp and set a high temp, while it SMBs, if you start rising toward the tail end of your COB.
katie disimone
@Kdisimone
May 09 2017 05:00
New SMB-pump-loop is not getting a move on…tried runagain.
Starting supermicrobolus pump-loop at Mon May 8 21:58:03 PDT 2017 with 24 second wait_for_silence:
Waiting up to 4 minutes for new BG: /usr/local/bin/oref0-pump-loop: line 383: jq: command not found
date: invalid date ‘@’
ls: cannot access monitor/pump_loop_completed: No such file or directory
Radio ok. Listening: .No pump comms detected from other rigs
Preflight OK. Profile less than 60m old. /usr/local/bin/oref0-pump-loop: line 322: jq: command not found
date: invalid date ‘@’
find: `monitor/pump_loop_completed': No such file or directory
pump_loop_completed less than 5m ago. RefreshRemoved duplicate bolus/carb entries:2
ed pumphistory
pumphistory.json: /usr/local/bin/oref0-pump-loop: line 75: jq: command not found
Error, retrying
been grinding on the same error message on repeat for well over 30 min.
sanderlong
@sanderlong
May 09 2017 05:10
Hi. I'm troubleshooting the oref0-setup.sh script. I've rerun all the steps 3 times and every time I get at message at the end that says: Checking bluez installation
cp: cannot stat ‘/root/src/openxshareble/bluetoothd.conf’: No such file or directory
Couldn't copy bluetoothd.conf
I'm running the script without BT tethering.
I've tried searching troubleshooting documents, but can't find anything similar
Dana Lewis
@danamlewis
May 09 2017 05:13
@sanderlong not sure anyone's troubleshooted the openxshareble option (nor has anyone done the work to get it done)
So best bet is to redo selecting G4 as the option if you want to loop straight away
@Kdisimone what did it do when you did a manual history pull?
sanderlong
@sanderlong
May 09 2017 05:15
Will try that. Thanks, Dana!
katie disimone
@Kdisimone
May 09 2017 05:15
@danamlewis i got crickets. it took the command but then went right back to the prompt. wondering if i did that wrong.
killall -g openaps
openaps refresh-old-pumphistory
do those sound like right commands?
I just tried new installation…no dice either. same error messages
katie disimone
@Kdisimone
May 09 2017 05:22
just tried a couple more...
Screen Shot 2017-05-08 at 10.21.20 PM.png
nope. same errors. argh
Tim Street
@tim2000s
May 09 2017 05:32
@Kdisimone For ore1 you need killall -g oref0-pump-Loop
Or maybe append the .sh on that.
katie disimone
@Kdisimone
May 09 2017 05:37
thanks Tim….i’m starting build from scratch. removed src and myopenaps. if the errors happen again…will do your correction.
Tim Street
@tim2000s
May 09 2017 05:37
Also what data source are you using? xdripaps?
katie disimone
@Kdisimone
May 09 2017 05:38
g4
Tim Street
@tim2000s
May 09 2017 05:39
Okay. Have no experience of that, but I see that error when my xdripaps raw data is empty and thebcomparison script isn't running properly.
katie disimone
@Kdisimone
May 09 2017 05:40
I wonder…drat. Wish I hadn’t wiped yet. I wonder if my flag was g4 and not g4-upload somehow
i was on master, switched to dev…i have a feeling that may have been it
katie disimone
@Kdisimone
May 09 2017 05:52
well…seeing same error again…i’ll wait a few minutes and then try manual
looks like it may have fixed...
at least got old pump history…that’s promising
sanderlong
@sanderlong
May 09 2017 05:53
Does anyone know why you would not be able to access the Edison after a reboot? I just set up papertrail, rebooted and now when I run ssh root@edisonhost.local, I get this error: ssh: Could not resolve hostname edisonhost.local: nodename nor servname provided, or not known
katie disimone
@Kdisimone
May 09 2017 05:53
working. i have a feeling it was the g4-upload
@sanderlong that is either when your rig is not on the same network as your computer, or your rig is still booting up and hasn’t gotten an IP address from your network quite yet, or your hostname (edisonhost) isn’t right
this page has common error messages and reasons when logging in http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-2/accessing-your-rig.html
if it’s just rebooted, give it a minute and try again
sanderlong
@sanderlong
May 09 2017 05:55
It was over 5 minutes ago
katie disimone
@Kdisimone
May 09 2017 05:56
oh, that’s definitely too long. Is your rig actually named edisonhost?
sanderlong
@sanderlong
May 09 2017 05:56
yes, I didn't rename it
katie disimone
@Kdisimone
May 09 2017 05:56
using terminal app?
sanderlong
@sanderlong
May 09 2017 05:56
It is actually plugged in to my computer
Yes, terminal
katie disimone
@Kdisimone
May 09 2017 05:57
Ok, then double check that your rig is plugged in through the UART port and then login using screen mode.
sudo screen /dev/tty.usbserial-* 115200
sanderlong
@sanderlong
May 09 2017 05:58
Sorry, I'm a little slow with this - what to I enter and where?
Rig is plugged in and the lights are on
katie disimone
@Kdisimone
May 09 2017 05:59
open a new terminal window. there’s two plug ports on your explorer board…make sure your cable is plugged into the UART port. Then copy and paste that line I posted above into the terminal prompt
sanderlong
@sanderlong
May 09 2017 05:59
Paste sudo screen /dev/tty.usbserial-* 115200 in terminal?
katie disimone
@Kdisimone
May 09 2017 05:59
yes
(The step-by-step is on this page…by the way. a good read)
then enter your computer’s password and press return a couple times
Scott Leibrand
@scottleibrand
May 09 2017 06:04
@Kdisimone you were missing jq originally.
sanderlong
@sanderlong
May 09 2017 06:04
Thanks. Unfortunately that didn't work either so I am going to have to restart
katie disimone
@Kdisimone
May 09 2017 06:04
@sanderlong if you read through that page, the steps to solve the errors are in there….restart is a common one for screen mode. :+1:
@scottleibrand how do i end up missing jq? I had a working loop on there earlier...
Scott Leibrand
@scottleibrand
May 09 2017 06:11
You probably had an old school loop that didn't need it, and never ran the newer install scripts that install it?
sanderlong
@sanderlong
May 09 2017 06:11
@Kdisimone, yes the restart worked. Thank you!
katie disimone
@Kdisimone
May 09 2017 06:13
Did oref0-setup.sh get updated to include that?
I’m trying to think back to how I updated that rig. I think I changed branches, updated, did the npm-global install. Then I’m a little fuzzy on how I ended up updating after that. I think I may have never run the whole setup script again and instead did run again only (with the microbolus flag added). Well, I’ve got 2 other rigs to try to recreate the error. :smile:
fidoman
@efidoman
May 09 2017 10:16
@scottleibrand - I tried testing it this morning, but must have checked out the branch wrong. I get a syntax error: /usr/local/bin/oref0-pump-loop: line 222: syntax error near unexpected token `fi'
@scottleibrand - what is the git command I should use to make sure I am using the #462 pull request?
dmdfreak
@dmdfreak
May 09 2017 12:14
pumphistory.json: "Bolus 2017-05-09T02:36:08 head[4], body[0] op[0x01]"
Checking pump clock: "2017-05-09T07:10:35-05:00" is within 1m of current time: Tue May  9 07:10:52 CDT 2017
and that pumphistory is less than 1m old.  Temp refreshed
{"carbs":0,"boluses":0,"mealCOB":0,"currentDeviation":null,"maxDeviation":0,"minDeviationSlope":0}
BG data is too old (it's probably this), or clock set incorrectly.  The last BG data was read at Tue May 09 2017 00:52:05 GMT-0500 (CDT) but your system time currently is Tue May 09 2017 07:11:04 GMT-0500 (CDT)
Checking deliverAt: null is within 1m of current time: Tue May  9 07:11:05 CDT 2017
date: invalid date ‘null’
Error, retrying
Waiting for 30s silence before mmtuning
Tried rerunning oref0 but I'm getting the same error
Also, NS stopped getting my numbers in the middle of the night. Did I see there was a way to get those many hours of numbers back into NS? They are there in xdrip+
dmdfreak
@dmdfreak
May 09 2017 12:34
The above error happened at 1am after I had been sleeping for a few hours. It was working fine when I went to sleep. I thought there was talk about it yesterday on here, but I haven't been able to locate it. Any ideas??
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 12:59
@dmdfreak there was talk about it yesterday
check posts by @scottleibrand and @danamlewis I think @scottleibrand posted something let me look in history around 12 pm tomorrow and see what I can find
David
@davidkeddydb
May 09 2017 13:02
Screenshot 2017-05-09 09.01.04.png
@tepidjuice @scottleibrand I still have to update my NS site to the newest, I just have to do that. I wish I could alter the lower temping that SMB is doing.
I wonder if I had UAM on if it would think I had floating carbs going and jumped in sooner. I know my carb count, however was dead on
dmdfreak
@dmdfreak
May 09 2017 13:07
thanks @diabetic_dude_twitter I was lurking during all that last night here and there , but at the time it didn't affect me and I had no help. But this morning I went back to try to find it and the closest thing I saw was someone had invalid date '@' or something like that. I was unsure if it was the same exact thing.
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 13:11
@dmdfreak are you using the dev or master version of 0ref0?
@dmdfreak this is what was posted by Scott Leibrand "regarding pump clock mismatches, I just pushed https://github.com/openaps/oref0/tree/set-pump-clock if anyone wants to test it (by setting your pump clock to >10s off, and verifying it re-sets it). haven't tested it yet myself, as it's bed time."
but honestly, if it is working today, I would just watch it and see if it happens again .
dmdfreak
@dmdfreak
May 09 2017 13:20
I'm on dev
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 13:21
yeah I would look into that link I posted, but wait for more issues to come up
when was the last time you updated your nightscout site ?
dmdfreak
@dmdfreak
May 09 2017 13:23
I've got everything new running. SMB and everything was great until I went to bed and woke up with this error. I'm just still gun shy on github merging. I'm about to give it a shot though.
floating carbs on NS included
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 13:29
I just did the NS merge
if you need help
(sorry that is why it was silent, I was reading how to do it) @dmdfreak
tell me if you need the link to check and start the nightscout update
katie disimone
@Kdisimone
May 09 2017 13:35
@dmdfreak that was my error. I solved my problem using the nuclear option (deleting my src and myopenaps and running phase 2 over again). But I'll put more effort into updating the next rig to really troubleshoot better. It's just that starting from scratch on phase 2 really doesn't take that long so I did that
@dmdfreak does have the proper NS branch and update. 👍🏻 I do remember that from just the other day
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 13:37
heck I prefer the nuclear option sometimes, what is what I am coming back up from, but I started from the very beginning and even re-flashed jubilinux
was having time issues also but a different issue, I think mine was related to my work's NTP restrictions
had to kill the bluetooth and all wireless configs and did not want to miss anything so just started from square 1 and getting very good success
Scott Leibrand
@scottleibrand
May 09 2017 13:39
@Kdisimone looks like oref0-setup only installs jq if you select autotune or pancreabble. this might be one you could fix with a PR to oref0-setup, if you'd like
you'd just need to add jq to the end of the apt-get install bc line for microbolus, so it does sudo apt-get -y install bc jq
katie disimone
@Kdisimone
May 09 2017 13:42
@scottleibrand i can help with that
(Although as side note, the other rig that I did from fresh flash forward never had autotune installed and has worked.)
I wonder if oref0-setup is ok but runagain was the problem?
Scott Leibrand
@scottleibrand
May 09 2017 13:46
@ericfeibelman or, you know, I could have made a stupid syntax error and forgot to actually run my modified script to check for that. ;-) fixed
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 13:48
programmers never make errors like that, it must have been a keyboard issue
katie disimone
@Kdisimone
May 09 2017 13:48
It's possible I did the whole update using runagain and editing that script vs using setup script over (although my memory was that I couldn't get it to work using runagain and had to use setup script instead...could be wrong. I have two rigs that I'll update tonight and take better notes. Is there a line command to check if those have jq installed already?)
Scott Leibrand
@scottleibrand
May 09 2017 13:49
@diabetic_dude_twitter it's just a small matter of typing: what's the problem? ;-)
@davidkeddydb @tepidjuice do you run autotune? I think your carb ratio may be too weak.
@Kdisimone this jq thing is a corner case that only affects old rigs that didn't get jq installed in the updated step 0/1 stuff (because they were done before that was added there)
katie disimone
@Kdisimone
May 09 2017 13:51
I can find those corner cases so well 🤣
Scott Leibrand
@scottleibrand
May 09 2017 13:54
openaps/docs@15a2362 is the commit where we added jq to quick-packages 2 months ago
David
@davidkeddydb
May 09 2017 13:54
@scottleibrand Yes, I have autotune running on this rig
Scott Leibrand
@scottleibrand
May 09 2017 13:54
@kdisimone and of course it only affects you because you don't run autotune. so yeah. ;-)
David
@davidkeddydb
May 09 2017 13:56
@scottleibrand I used the bolus wizard on the pump though. let me see what autotune would have done
live4sw
@live4sw
May 09 2017 13:57
@mhaeberli FYI - after working fine last night, unfortunately the time error problem is recurring with my Pi rig
blob
Scott Leibrand
@scottleibrand
May 09 2017 13:58
@davidkeddydb that may or may not be a major contributor to your problem. the other underlying issue is that we designed SMB's behavior to work best when you get insulin activity going before the meal via an Eating Soon or by pre-entering carbs.
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 13:59
@live4sw does that Pi rig move around with you or stay in one location?
David
@davidkeddydb
May 09 2017 13:59
@scottleibrand I did do eating soon and I prebolused though only like 10 minutes prior to eating
@scottleibrand I was I think 30 minutes prior with eating soon
live4sw
@live4sw
May 09 2017 14:00
@diabetic_dude_twitter Moves around, this is my mobile rig. The other two at home are working fine. All three are Pi3's with Ti Stick, set up using the most recent setup script.
Scott Leibrand
@scottleibrand
May 09 2017 14:00
k. based on your graph, I'm skeptical that the zero-temping was the main cause of the meal rise. it didn't help though.
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:01
@live4sw are you on dev?
live4sw
@live4sw
May 09 2017 14:01
Though for odd reasons I can't figure out, when I run the curl command to install master, at the end my two home rigs return "openaps 0.1.5" while the mobile one returns "openaps 0.2.0-dev".
Scott Leibrand
@scottleibrand
May 09 2017 14:01
if you had previously installed the newer version it won't downgrade it
live4sw
@live4sw
May 09 2017 14:02
Perhaps I messed up at some point and installed dev on the mobile one by mistake. Is there a known clock issue with dev not present in master?
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:03
I can not speak for you but I was having the same problems you are having for 4 weeks
I finally reflashed to fix my issues
then only installed wireless for my phone and home
going to add things back slowly to trace where the problem was coming from
live4sw
@live4sw
May 09 2017 14:04
You were on dev?
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:04
NO
I want to fix this before creating another variable that could contribute to problems
Scott Leibrand
@scottleibrand
May 09 2017 14:05
@davidkeddydb we would expect to see the impact of the zero temp to peak about 75m afterward. so since the zero temping was at 8am +/- 30m, we would expect its affect to be at about 9:15am +/- 45m or so. however, your meal rise is mostly in the 8:30am +/- 30m or so timeframe, which would be dominated by the effects of the meal bolus itself, not the subsequent zero temps
dmdfreak
@dmdfreak
May 09 2017 14:05
Thanks again @Kdisimone I was holding off on the nuclear option. how do I delete those and to clarify, by phase 2 you just mean the latest oref0 with SMB correct?
live4sw
@live4sw
May 09 2017 14:06
Ah. Well it seems reasonable to me that I should set up this rig again from scratch. Had sorta been avoiding it but now that we have the setup script, setting up new rigs is so much easier than before.
David
@davidkeddydb
May 09 2017 14:07
@scottleibrand :+1:
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:07
@live4sw I would see about backing up your WPA_supplicant.conf and your settings in oref0 to make the reinstall easier
(that is what I did)
what I am doing now is adding the info from the WPA_supplicant.conf network by network until I get the error again.
if I get the error I will tell you when it comes in. I have already had problem with my works wifi so I removed both of those networks
I think the firewalls here are blocking external NTP communications
(even though I am able to ping them)
live4sw
@live4sw
May 09 2017 14:09
that makes a lot of sense too. when i was checking the logs it would say something like "resetting NTP, not ok"
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:10
yeah I was getting that same error
live4sw
@live4sw
May 09 2017 14:10
and I only noticed that in the logs when i was at work
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:10
@live4sw this is a lot of the troubleshooting I have done https://www.imzy.com/openaps/post/openaps_does_not_seem_to_be_communicating_with_my_
the info will not help you much except for you to confirm what you also have done and the same errors
live4sw
@live4sw
May 09 2017 14:12
looks very familiar
katie disimone
@Kdisimone
May 09 2017 14:13

@dmdfreak rm -rf myopenaps and rm -rf src will delete your loop and oref0. Then you start at step 0 of Phase 2, install dependencies. (All your wifi stuff will be fine). That's the nuclear option

http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-2/oref0-setup.html

Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:13
@live4sw my main fix has been the reflash and then moving slowly and checking my networks, but it has paid off, like I said, two networks so far are bad and I have only added 4 that is a 50% failure chance
Scott Leibrand
@scottleibrand
May 09 2017 14:14
@kdisimone that sounds more like the bulldozer option: it leaves the jubilinux foundation in place. the nuclear option would be more like re-flashing. ;-)
katie disimone
@Kdisimone
May 09 2017 14:14
@kenstack had a way around blocked ntp but I don't know the details
@scottleibrand good point. It does preserve a fair bit of work still.
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:15
@Kdisimone correct he gets the time from the pump
I am waiting on a response on that one
dmdfreak
@dmdfreak
May 09 2017 14:15
I think my nuclear option a week ago was go all the way back and redo everything including the OS steps. I figured I was over doing it but it worked. Didn't want to go that route again this time. Thats why I was waiting for an answer before hitting the button. Thank you so much. Launch in progress.....
Scott Leibrand
@scottleibrand
May 09 2017 14:15
maybe I should go ahead and merge my PR that sets time properly from the pump for SMB... :-)
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:16
if it works, please do so sir
dmdfreak
@dmdfreak
May 09 2017 14:16
that a hint to wait a minute? LOL
Scott Leibrand
@scottleibrand
May 09 2017 14:16
openaps/oref0#460 merged to dev
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:17
woohoo! one less issue when I update (a long itme from now lol )
Scott Leibrand
@scottleibrand
May 09 2017 14:17
FWIW, all of y'all can read and comment on the PR's at https://github.com/openaps/oref0/pulls and give a :+1: for any that look good and you'd like merged faster.
often I end up waiting longer than necessary to merge things because no one is giving feedback on them.
live4sw
@live4sw
May 09 2017 14:18
@diabetic_dude_twitter Thanks, sounds like step 1 will be run curl again on the rig to pick up Scott's newest addition, step 2 would be remove work network, step 3 will be start from scratch.
Scott Leibrand
@scottleibrand
May 09 2017 14:18
you'll just need to git pull in ~/src/oref0/ to pull in newly merged PRs
live4sw
@live4sw
May 09 2017 14:19
oh ok
Scott Leibrand
@scottleibrand
May 09 2017 14:19
if they're major changes that add new commands, you sometimes have to do npm run global-install there again as well, but not for something minor like this
David
@davidkeddydb
May 09 2017 14:31
Anyone know what my NS site should say for version if I correctly got the floating-carbs merged right?
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 14:33
@davidkeddydb I am not sure in your case but you can check if you are at the latest version by going here: http://nightscout.github.io/pages/update-fork/
David
@davidkeddydb
May 09 2017 14:34
@diabetic_dude_twitter I screwed up and went to dev rather than floating-carbs so the fork update says I am updated, but not to what I want to have
@diabetic_dude_twitter Think I got it. version 0.10.0-dev-20170423 hopefully is right
Scott Leibrand
@scottleibrand
May 09 2017 14:44
I neglected to update the version number on the NS floating-carbs branch.
Eric
@ecc1
May 09 2017 14:54
@diabetic_dude_twitter You can verify whether your firewall is blocking NTP by using the ntpdate command with a specific NTP server as argument. It's in a separate package so you'd have to apt-get install it first.
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 15:00
thanks @ecc1
live4sw
@live4sw
May 09 2017 15:01
Random question - I can't seem to connect to my pi via SSH in the office using its hostname. Is there a way to figure out its current IP using either my Heroku site or my papertrail logs?
Eric
@ecc1
May 09 2017 15:04
Each papertrail log line has the IP address of the source (2nd field, after the timestamp)
You can also look at the papertrail dashboard, and click on "All Systems"; it will give you a pie chart of IP addresses in use
live4sw
@live4sw
May 09 2017 15:22
That is very interesting. It looks like because I am behind a work network, it is presenting my IP from my work computer and my rig as the same IP. I'm no networking expert, but is that some sort of common security mechanism?
Eric
@ecc1
May 09 2017 15:25
Not so much security as conservation of scarce IPv4 addresses. It's called NAT; only your workplace router has an externally routable IP address, and takes care of all the private IP addresses of computers "downstream" from it. That way you can use 192.168.x.x for your local machines and so can I, without conflict.
live4sw
@live4sw
May 09 2017 15:26
Ahh. So that explains why I'm unable to figure out the IP address of my rig when it's behind the workplace network
It also tends to reinforce the notion that it is my workplace network that is preventing my rig from updating time using NTP, and the way to test that hypothesis would be to remove my workplace network from wpa_supplicant
Eric
@ecc1
May 09 2017 15:28
Right. It also usually means that the only way to reach a machine behind a NAT firewall is if it reaches you first; i.e. by responding to it, rather than initiating a connection. There are various ways around this, usually involving a third-party site as intermediary.
live4sw
@live4sw
May 09 2017 15:31
Ah. Would the same be true if I were tethering, it would just present the phone hotspot's IP to the outside world when it was actually my rig connecting?
Scott Leibrand
@scottleibrand
May 09 2017 15:31
Yes, all of your IPv4 connectivity will be through a NAT of some sort.
Eric
@ecc1
May 09 2017 15:31
I haven't confirmed it, but almost certainly.
live4sw
@live4sw
May 09 2017 15:33
Got it. That's very useful, frustrating that I can't figure out the IP address solely from the logs so that I can fix it now, but at least seems to shed light that it was a network connectivity problem that is causing the time mismatch error that @mhaeberli and I have been experiencing.
Eric
@ecc1
May 09 2017 15:33
But note that you can do something as simple as log a message (to papertrail!) or email yourself a note containing your IP addresses each time your rig boots (or even changes IP address, if you want to get fancier with /etc/network/if-up.d/ rules
Scott Leibrand
@scottleibrand
May 09 2017 15:33
If you look at /var/log/openaps/network.log, it has both your private and public IP listed.
live4sw
@live4sw
May 09 2017 15:35
That's really helpful (unfortunately can't get onto the rig until tonight, but still). Contemplating going out of range of my work wifi so it connects to hotspot instead. For purposes of SSH, would I then be able to connect from the hotspot phone to the rig using its hostname as if they were on the same network?
That would then let me clear my work wifi from wpa_supplicant so it will just use hotspot from now on.
Scott Leibrand
@scottleibrand
May 09 2017 15:37
If you look at /var/log/openaps/network.log, it has both your private and public IP listed. Or you can use the manual methods in the docs for figuring it out and accessing your rig, perhaps over a UART console link. (Sorry for any dupes. Out hiking with limited signal.)
live4sw
@live4sw
May 09 2017 15:37
Not a problem, you guys are a huge help!
Scott Leibrand
@scottleibrand
May 09 2017 15:37
Papertrail might be able to report the contents of network.log too.
We also put our phone wifi hotspot into wpa_supplicant with a higher priority so we can force it off other wifi without walking away.
Martin Haeberli
@mhaeberli
May 09 2017 15:39
so - having done a git pull in oref0 - what obvious magic if any is needed to re-build to benefit from updates? something like killall -g openaps; xxx; (relaunch cron jobs)
Scott Leibrand
@scottleibrand
May 09 2017 15:40
Depends on what was in the git pull
Martin Haeberli
@mhaeberli
May 09 2017 15:40
ok, but most conservatively (that is, most work)? - say, assume your clock fix
which is probably NOT a major change
Scott Leibrand
@scottleibrand
May 09 2017 15:41
An npm run global-install is needed for major changes, but not that. Just need to restart oref0-pump-loop or let it complete and start again.
Martin Haeberli
@mhaeberli
May 09 2017 15:41
:+1:
Scott Leibrand
@scottleibrand
May 09 2017 15:41
With SMB the killall needs to be killall -g oref0-pump-loop instead.
Martin Haeberli
@mhaeberli
May 09 2017 15:41
:+1:
Scott Leibrand
@scottleibrand
May 09 2017 15:41
(If you're impatient.)
elodaille01
@elodaille01
May 09 2017 15:56
Hi ! is there any variable to add in Heroku to see what SMB is doing on NS @scottleibrand please ? ;) :)
Dana Lewis
@danamlewis
May 09 2017 15:59
@elodaille01 switch to the floating carbs branch of NS for additional purple line for UAM. Otherwise, it's all in the OpenAPS pill and the regular logs (plus boluses will show on the NS graph)
elodaille01
@elodaille01
May 09 2017 16:00
Hi Dana..I already have the purple lines on my NS, does it mean that my version is ok ? Thank you Dana
Dana Lewis
@danamlewis
May 09 2017 16:01
No. See above. If you have UAM you will have 4 purple lines total if you get the floating carbs branch of Ns
elodaille01
@elodaille01
May 09 2017 16:01
Ok thank you, I just have 3 of them
kenstack
@kenstack
May 09 2017 16:16
@diabetic_dude_twitter @Kdisimone there are 2 options - easiest is if you have access to the internet but no ntp - then just grab time from google like so: sudo date -s "$(wget -qSO- --max-redirect=0 google.com 2>&1 | grep Date: | cut -d' ' -f5-8)Z"
if not, you have t read time from the pump. Im not sure what the openaps command is, but the correct tweak to decocare is ReadRTC - I can extract it from my code and send it later tonight Im not near my computer.
Dana Lewis
@danamlewis
May 09 2017 16:18
@elodaille01 I don't use heroku so I'd ask here so others can help troubleshoot if you are still stuck
elodaille01
@elodaille01
May 09 2017 16:19
ok Dana..thank you..
kenstack
@kenstack
May 09 2017 16:19
@scottleibrand or @danamlewis probably knows the openaps command to grab the time from the RTC of the pump. then use use date -s $time where time is the string you grabbed from the pump RTC. Again if that doesnt already exist I can extract it from my node code
elodaille01
@elodaille01
May 09 2017 16:26
Hi everybody, please if it's like this (see below) on my heroku dashboard (deploy), why is my NS not updated itself ?
blob
I'm still on 0.9.2..
Scott Leibrand
@scottleibrand
May 09 2017 16:29
@kenstack @diabetic_dude_twitter no need to do it manually: oref0 dev now runs oref0-set-system-clock as needed.
elodaille01
@elodaille01
May 09 2017 16:30
..I finally choose to deploy dev branch manually..We'll see
Scott Leibrand
@scottleibrand
May 09 2017 16:33
Master branch has done that for months: I just hadn't gotten it working with SMB until this morning's merge.
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 16:34
thanks for that info @scottleibrand
and @kenstack
I may do the bulldozer method and start over with oref0 dev and check out all that is going on with SMB
elodaille01
@elodaille01
May 09 2017 16:40
@Dana or @scottleibrand please, what is the version of NS including floating carbs ?
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 16:43
@elodaille01 you may want to include @davidkeddydb he was having a question like that this morning. looking in history to see what he found
elodaille01
@elodaille01
May 09 2017 16:44
ok thank you @diabetic_dude_twitter !
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 16:49
@elodaille01 from 9:41 AM "Think I got it. version 0.10.0-dev-20170423 hopefully is right"
he was on 0.9.somehting dev from reading earlier but I would guess (like him) that is the right version
elodaille01
@elodaille01
May 09 2017 16:51
ok thank you @diabetic_dude_twitter !
Scott Leibrand
@scottleibrand
May 09 2017 16:52
Don't pay as much attention to the version: make sure you deploy the NS floating-carbs branch if you wan the UAM purple line.
Also, if you're using that successfully, please :+1: the NS PR so we can get it merged to NS dev.
elodaille01
@elodaille01
May 09 2017 16:53
ok no problem @scottleibrand
dmdfreak
@dmdfreak
May 09 2017 17:17
@Kdisimone Nuclear Option installed but on reboot I got this. Waited to see if it would resolve while I did an errand and still an issue.
pumphistory.json: "Bolus 2017-05-09T10:19:12 head[4], body[0] op[0x01]"
Checking pump clock: "2017-05-09T12:10:33-05:00" is within 1m of current time: T                                                                                        ue May  9 12:10:49 CDT 2017
and that pumphistory is less than 1m old.  Temp refreshed
Could not parse input data:  [SyntaxError: /root/myopenaps/monitor/glucose.json:                                                                                         Unexpected end of input]
Traceback (most recent call last):
  File "/usr/local/bin/openaps-report", line 82, in <module>
    app( )
  File "/usr/local/lib/python2.7/dist-packages/openaps/cli/__init__.py", line 51                                                                                        , in __call__
    self.run(self.args)
  File "/usr/local/bin/openaps-report", line 75, in run
    output = app(args, self)
  File "/usr/local/lib/python2.7/dist-packages/openaps/cli/subcommand.py", line                                                                                         52, in __call__
    return self.method.main(args, app)
  File "/usr/local/lib/python2.7/dist-packages/openaps/reports/invoke.py", line                                                                                         48, in main
    reporters.Reporter(report, device, task)(output)
  File "/usr/local/lib/python2.7/dist-packages/openaps/reports/reporters/__init_                                                                                        _.py", line 33, in __call__
    self.blob = self.serialize(data)
  File "/usr/local/lib/python2.7/dist-packages/openaps/reports/reporters/__init_                                                                                        _.py", line 31, in serialize
    return self.method.serialize(render(data), self)
  File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/process.py", line                                                                                         56, in prerender_json
    return json.loads(data)
  File "/usr/lib/python2.7/json/__init__.py", line 338, in loads
    return _default_decoder.decode(s)
  File "/usr/lib/python2.7/json/decoder.py", line 366, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
  File "/usr/lib/python2.7/json/decoder.py", line 384, in raw_decode
    raise ValueError("No JSON object could be decoded")
ValueError: No JSON object could be decoded
Error, retrying
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 ri                                                                                        gs
mmtune: "916.588", 5, -78 waiting for 36 second silence before continuing
Radio ok. Listening: .No pump comms detected from other rigs
Sleeping 5;

Starting supermicrobolus pump-loop at Tue May 9 12:13:27 CDT 2017 with 17 second                                                                                         wait_for_silence:
Waiting up to 4 minutes for new BG: date: invalid date ‘@’
ls: cannot access monitor/pump_loop_completed: No such file or directory
Radio ok. Listening: .No pump comms detected from other rigs
Preflight OK. Profile less than 60m old. date: invalid date ‘@’
find: `monitor/pump_loop_completed': No such file or directory
pump_loop_completed less than 5m ago. Refresh    at node.js:906:3
ed pumphistory
katie disimone
@Kdisimone
May 09 2017 17:19
if it fails, you could try the manual installation of jq like @scottleibrand suggested sudo apt-get -y install bc jq
dmdfreak
@dmdfreak
May 09 2017 17:23
I ran that. Waiting. What is jq exactly?
ok. I'm going to nuck again. Maybe I got ahead of the clock fix.
katie disimone
@Kdisimone
May 09 2017 17:26
@scottleibrand can i add to the list in oref0/bin/oref0-log-shortcuts.sh to add (as a PR) the other aliases that I have here ( below) currently? I manually added them to my rig and work great, but I’d love to add them to the basic install so that others had aliases like edit-pref available http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-1/add-alias.html
then we could delete that page from the docs when they are simply part of that script
and maybe change the networklog and autotunelog aliases to network-looplog and autotune-looplog just to be consistent with the other aliases.? I won’t remember that the those two are the ones that are actually named differently for the log…so if they all have the same -extension for the alias, I’m more likely to remember them vs. having to look them up.
Scott Leibrand
@scottleibrand
May 09 2017 17:33
Yes please.
katie disimone
@Kdisimone
May 09 2017 17:35
:+1:
katie disimone
@Kdisimone
May 09 2017 17:47
@scottleibrand is it possible to have an alias for a bash command? Like grep -q runagain ~/.bash_profile || echo "alias runagain="'"bash ~/myopenaps/oref0-runagain.sh"' >> ~/.bash_profile
if that doesn’t look like it would break, i’d like to add that since that’s one I still look up and copy/paste a lot
Dana Lewis
@danamlewis
May 09 2017 17:53
#addallthethings :smile:
aeandrews
@aeandrews
May 09 2017 17:56
Hi all, I'm building an RPi3 rig with a carelink and attempting to communicate with it using terminal on my Mac. Last night was my first time successfully setting up the loop and letting it run automated overnight, which worked well until the pump and the rig stopped communicating at some point while I was asleep. I'm now having trouble communicating with my rig via terminal - when I try to ssh into the pi, I keep getting the error "Could not resolve hostname raspberrypi.local: nodename nor servname provided, or not known." Both my rig and computer are on my home wifi network. I also tried changing the battery in my pump since I'd gotten some battery alerts last night before I went to sleep, and I was wondering if that might have been the reason why the rig and pump stopped communicating, but it doesn't explain why I can't get into the rig now. I'm trying to follow this page in the docs (http://openaps.readthedocs.io/en/latest/docs/walkthrough/phase-2/accessing-your-rig.html?highlight=hostname) but the explanation for the error there says that my rig and computer are probably on different networks, which I know isn't true. Any help would be much appreciated, thanks :)
katie disimone
@Kdisimone
May 09 2017 17:57
@aeandrews can you confirm that your rig is connected? Try logging onto your router? usually you can do that by going to the URL 192.168.1.1
dmdfreak
@dmdfreak
May 09 2017 18:00
Still the same problem..
xdrip isn't updating NS either
jmcrawford
@jmcrawford
May 09 2017 18:06
@dmdfreak how long has xdrip not been uploading? Mine stopped working about an hour ago
aeandrews
@aeandrews
May 09 2017 18:08
@Kdisimone thanks for the tip. i logged in and it seems my rig isn't connected (it's not showing up in the list of wireless clients). should i go back to an earlier point in the docs and reconfigure the wifi?
katie disimone
@Kdisimone
May 09 2017 18:09
I’m not super well versed in rpi setups. For us on explorer, we’d try a reboot…and if that doesn’t work, we’d login using screen mode and a cable
Jason Wittmer
@jbwittmer
May 09 2017 18:09
@aeandrews If you set up on a Raspberry Pi using the script and are NOT using bluetooth, you may be encountering an error that others of us have noticed and is referenced here: openaps/oref0#371 I have found by disabling the oref-online ( put a # in front of it) in the crontab that it does it less frequently. If you can plug the Pi into an ethernet cable I think you'll find that you can ssh into the rig and make the changes that way, or you can just risk it and unplug and plug in the rig again (although there is some risk of corrupting your SD card).
katie disimone
@Kdisimone
May 09 2017 18:09
I doubt your wifi files need to be adjusted if they worked before
(ah! thanks @jbwittmer )
aeandrews
@aeandrews
May 09 2017 18:11
@jbwittmer @Kdisimone i'm not using bluetooth so that may be it! I'll give that a shot. thanks to you both!
Jason Wittmer
@jbwittmer
May 09 2017 18:12
I will say that I still run into this issue occasionally even with the above fix, but it happens a lot less often.
dmdfreak
@dmdfreak
May 09 2017 18:14
@jmcrawford about 1am last night while I was sleeping. Running fine for hours then the rig threw an error and stopped working right and xdrip stopped updating
Sgoldberg89
@Sgoldberg89
May 09 2017 18:15
@scottleibrand My explorer board has had the same issue again with the subg_rfspy. Lasted about 24 hours before it went out again. Is this a known issue?
Scott Leibrand
@scottleibrand
May 09 2017 18:20
That has aged out of my mental cache already. What issue?
Sgoldberg89
@Sgoldberg89
May 09 2017 18:20
We had to reflash the subg_rfspy stuff?
and did the PR on the page with instructions on how to do it
jmcrawford
@jmcrawford
May 09 2017 18:20
@dmdfreak There was a post on CGMitC with the same issue, xDrip stopped puloading. But the post is not showing up now. No one else had commented on it yet.
xDrip just uploaded all the data from the last hour - not sure what happened.
dmdfreak
@dmdfreak
May 09 2017 18:30
No luck for me but I'm glad yours uploaded. I think my issue is different. Something with my rig
Think I'm going back to phase 1 restart when I get back home. I feel lost without my openaps. Amazing how fast you get used to it hehe
Martin Haeberli
@mhaeberli
May 09 2017 18:32
@dmdfreak :-( in sympathy
dmdfreak
@dmdfreak
May 09 2017 18:33
I just wish I understood what happened.
Guess I could try a redo on master and see if it comes back first. Then if it does try Dev again
aeandrews
@aeandrews
May 09 2017 18:38
@jbwittmer i don't have an ethernet cord so just unplugged it and can now ssh in, yay! just checking, i should be changing this in crontab: ps aux | grep -v grep | grep -q "oref0-online " || oref0-online to ps aux | grep -v grep | grep -q "oref0-online " ||#oref0-online ? (i am not a programmer so i'm new to this!)
Jason Wittmer
@jbwittmer
May 09 2017 18:41
@aeandrews I'm not by our Pi rig right now, so I can't remember exactly what you'll see, but what I have done is put a hashtag (#) in front of that entire command line. That will disable the entire thing. Since you're not using Bluetooth, it shouldn't negatively affect the loop otherwise (at least it hasn't in my experience).
Scott Leibrand
@scottleibrand
May 09 2017 18:42
@Sgoldberg89 were the lights solid again? Did ccprog fix it again?
I haven't heard of any issues causing that repeatedly, but if it keeps happening it might be worth trying different hardware and see if the problem follows the Explorer board.
dmdfreak
@dmdfreak
May 09 2017 18:56
Would you happen to have any ideas @scottleibrand before blow it up?
Scott Leibrand
@scottleibrand
May 09 2017 18:58
Haven't been following. What's the issue?
dmdfreak
@dmdfreak
May 09 2017 18:58
I only ask because it was having line errors even on a phase 2 reinstall. I even deleted to sec and myooenaos dir
One sec I'll try to copy paste it again.
Mobile ight copy paste friendly
Src and myopenaps dir*
live4sw
@live4sw
May 09 2017 19:05
Useful tip for anyone trying to SSH to their mobile rig on-the-go and the hostname isn't working. Simply connect to your wireless hotspot using another device, use Fing or another network scanner to figure out the IP of your rig, then connect using a mobile SSH app. Was able to do a quick config fix on my rig that way.
Got this idea from "How to determine the IP address of your rig?" in the docs, super useful.
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 19:06
yup @live4sw I have been there
dmdfreak
@dmdfreak
May 09 2017 19:06

pumphistory.json: "Bolus 2017-05-09T10:19:12 head[4], body[0] op[0x01]"
Checking pump clock: "2017-05-09T12:10:33-05:00" is within 1m of current time: T ue May 9 12:10:49 CDT 2017
and that pumphistory is less than 1m old. Temp refreshed
Could not parse input data: [SyntaxError: /root/myopenaps/monitor/glucose.json: Unexpected end of input]
Traceback (most recent call last):
File "/usr/local/bin/openaps-report", line 82, in <module>
app( )
File "/usr/local/lib/python2.7/dist-packages/openaps/cli/init.py", line 51 , in call
self.run(self.args)
File "/usr/local/bin/openaps-report", line 75, in run
output = app(args, self)
File "/usr/local/lib/python2.7/dist-packages/openaps/cli/subcommand.py", line 52, in call
return self.method.main(args, app)
File "/usr/local/lib/python2.7/dist-packages/openaps/reports/invoke.py", line 48, in main
reporters.Reporter(report, device, task)(output)
File "/usr/local/lib/python2.7/dist-packages/openaps/reports/reporters/init .py", line 33, in call
self.blob = self.serialize(data)
File "/usr/local/lib/python2.7/dist-packages/openaps/reports/reporters/
init .py", line 31, in serialize
return self.method.serialize(render(data), self)
File "/usr/local/lib/python2.7/dist-packages/openaps/vendors/process.py", line 56, in prerender_json
return json.loads(data)
File "/usr/lib/python2.7/json/init.py", line 338, in loads
return _default_decoder.decode(s)
File "/usr/lib/python2.7/json/decoder.py", line 366, in decode
obj, end = self.raw_decode(s, idx=_w(s, 0).end())
File "/usr/lib/python2.7/json/decoder.py", line 384, in raw_decode
raise ValueError("No JSON object could be decoded")
ValueError: No JSON object could be decoded
Error, retrying
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 ri gs
mmtune: "916.588", 5, -78 waiting for 36 second silence before continuing
Radio ok. Listening: .No pump comms detected from other rigs
Sleeping 5;

Starting supermicrobolus pump-loop at Tue May 9 12:13:27 CDT 2017 with 17 second wait_for_silence:
Waiting up to 4 minutes for new BG: date: invalid date ‘@’
ls: cannot access monitor/pump_loop_completed: No such file or directory
Radio ok. Listening: .No pump comms detected from other rigs
Preflight OK. Profile less than 60m old. date: invalid date ‘@’
find: `monitor/pump_loop_completed': No such file or directory
pump_loop_completed less than 5m ago. Refresh at node.js:906:3
ed pumphistory

Scott Leibrand
@scottleibrand
May 09 2017 19:07
@dmdfreak looks like your problem is this: Could not parse input data: [SyntaxError: /root/myopenaps/monitor/glucose.json: Unexpected end of input]
so for some reason you're not getting any glucose data onto your rig.
dmdfreak
@dmdfreak
May 09 2017 19:09
The rig stopped talking to the pump at the same time and xdrip stopped updating NS as well.
Scott Leibrand
@scottleibrand
May 09 2017 19:09
how do you normally get CGM data?
the rig is talking to the pump just fine: mmtune: "916.588", 5, -78
dmdfreak
@dmdfreak
May 09 2017 19:10
G5 to xdrip
Scott Leibrand
@scottleibrand
May 09 2017 19:10
and are you trying to use xdripaps with the rig, or uploading that to NS with the app?
dmdfreak
@dmdfreak
May 09 2017 19:10
Xdripaps
Scott Leibrand
@scottleibrand
May 09 2017 19:11
then I'm of no use: I've never used that and don't know how to troubleshoot it.
as a workaround you could try getting BGs through NS
dmdfreak
@dmdfreak
May 09 2017 19:11
So on install choose g5?
Scott Leibrand
@scottleibrand
May 09 2017 19:12
doesn't matter: regardless of what you pick it'll try to get BG both from that and from NS
so you can leave your current setup on the rig
the G5 setting is an option if you have a microUSB OTG adapter or an OTG male-male cable and want to plug in your receiver.
dmdfreak
@dmdfreak
May 09 2017 19:13
I already have mongo setup in xdrip and turned on. That's how it's always been set.
Oh ok
David
@davidkeddydb
May 09 2017 19:16
Screenshot 2017-05-09 15.05.37.png
So is this saying that autotune thinks my settings are correct for basals?
Dana Lewis
@danamlewis
May 09 2017 19:18
see how iSF and carb ratio is identical?
usually means something is buggy - no one’s isf and basals are spot on like that
@davidkeddydb how are you running autotune? one-off (phase c), one off as part of your rig, or automatically at night?
David
@davidkeddydb
May 09 2017 19:19
@danamlewis I just thought I nailed it. :) Autotune is running as part of the rig
running it automatically
[ { i: 0, start: '00:00:00', rate: 1.134, minutes: 0 },
{ i: 1, start: '01:00:00', rate: 1.114, minutes: 60 },
{ i: 2, start: '02:00:00', rate: 1.105, minutes: 120 },
{ i: 3, start: '03:00:00', rate: 0.973, minutes: 180 },
{ i: 4, start: '04:00:00', rate: 0.981, minutes: 240 },
{ i: 5, start: '05:00:00', rate: 1.314, minutes: 300 },
{ i: 6, start: '06:00:00', rate: 1.297, minutes: 360 },
{ i: 7, start: '07:00:00', rate: 1.286, minutes: 420 },
{ i: 8, start: '08:00:00', rate: 1.305, minutes: 480 },
{ i: 9, start: '09:00:00', rate: 0.587, minutes: 540 },
{ i: 10, start: '10:00:00', rate: 0.6, minutes: 600 },
{ i: 11, start: '11:00:00', rate: 0.6, minutes: 660 },
{ i: 12, start: '12:00:00', rate: 0.594, minutes: 720 },
{ i: 13, start: '13:00:00', rate: 1.373, minutes: 780 },
{ i: 14, start: '14:00:00', rate: 1.373, minutes: 840 },
{ i: 15, start: '15:00:00', rate: 1.09, minutes: 900 },
{ i: 16, start: '16:00:00', rate: 1.1, minutes: 960 },
{ i: 17, start: '17:00:00', rate: 1.1, minutes: 1020 },
{ i: 18, start: '18:00:00', rate: 1.056, minutes: 1080 },
{ i: 19, start: '19:00:00', rate: 1.071, minutes: 1140 },
{ i: 20, start: '20:00:00', rate: 0.965, minutes: 1200 },
{ i: 21, start: '21:00:00', rate: 0.994, minutes: 1260 },
{ i: 22, start: '22:00:00', rate: 0.976, minutes: 1320 },
{ i: 23, start: '23:00:00', rate: 0.989, minutes: 1380 } ]
p50deviation: -0.91 p50BGI -5.07 p50ratios: 1.162 Old ISF: 65 fullNewISF: 75.53 adjustedISF: 70.265 newISF: 65.527
this is from the daily today, well midnight:
Scott Leibrand
@scottleibrand
May 09 2017 19:23
that looks good. you'll need to figure out what happened after that so that the recommendations file didn't get updated
Chuck Vann
@djtechman
May 09 2017 19:25
can anyone point me in the right direction for help setting up SSL/https for my nightscout docker install?
ls
Eric
@ecc1
May 09 2017 19:28
There are a few ways, depending on how you're running the NS server. I'm using the "bare" nodejs server, configured with PORT=443, and using acmetool (let's encrypt everything) in redirector mode. Others use nginx or apache in front of NS, and corresponding let's encrypt setups.
Chuck Vann
@djtechman
May 09 2017 19:29
I just set things up per the nightscout-docker guide (defaults)
thanks for the recommendations, I'll search up on those methods
JELCRAWFORD
@JELCRAWFORD
May 09 2017 20:04
Openaps logic - may 9-17.png
I'm hoping someone can help us to understand the piece of the logic that includes the 0.217>2 in the above Openaps info. it has been appearing in this way periodically for a while and we don't really understand the meaning behind the info
Scott Leibrand
@scottleibrand
May 09 2017 20:06
the full statement is 13m@1.000 = 0.217 > 2 * req 0.06
that means "running 13m more at 1U/hr would result in 0.217U of extra insulin, which is more than double the required extra insulin of 0.06U"
if you have any ideas for making that clearer (without adding a bunch more words) it could definitely use some clarification/simplification
Dana Lewis
@danamlewis
May 09 2017 20:12
(2 times the required .06)
:+1:
katie disimone
@Kdisimone
May 09 2017 20:14
@JELCRAWFORD this has the logic (albeit in code speak for some of it…but there are code notes that give help to understand the logics…) I’m spending a lot of time on this page to make myself a cheat sheet to help follow logic flow. It’s pretty fun. Your particular example is down near line 475. https://github.com/openaps/oref0/blob/master/lib/determine-basal/determine-basal.js
JELCRAWFORD
@JELCRAWFORD
May 09 2017 20:14
Ok, so it is saying that because the (then) current basal was more than 2 higher, it is reducing the temp to 0.625 - is that correct. I'm assuming we will only see the "2 " logic when there is a reduction from pump programmed basal greater than times 2? We certainly don't see it all the time, but of course there is not always a reduction of size either.... not sure how to say it in minimum characters except to perhaps use an example in the documentation of a few outputs so that people can learn the various ways data will be expressed.
Ahh... I see @Kdisimone is already way ahead of me :)
I think that is usually the case
Scott Leibrand
@scottleibrand
May 09 2017 20:16
and yes, it's reducing the temp to 0.625U/hr. sometimes you'll get the 2 * req thing but then it'll say (in code-speak) "oh, never mind, what you've got is close enough" and not change it.
fidoman
@efidoman
May 09 2017 20:29

@scottleibrand - I'm fairly sure that this is breaking my loop ...

Checking system clock against pump clock:
Checking deliverAt: null is within 1m of current time: Tue May 9 16:27:34 EDT 2017
date: invalid date ‘null’
Error, retrying

Scott Leibrand
@scottleibrand
May 09 2017 20:31
that's probably a symptom of something else. have a meeting right now, but I'll be online later to troubleshoot further if you don't figure it out by then
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 20:33
what is the easiest way to "start over" I have a new NS site and want to use the dev build of openAPS... isnt there a command to remove the src folder on my edison and should I also remove my openaps folder?
Scott Leibrand
@scottleibrand
May 09 2017 20:35
@ericfeibelman maybe start with grep "MDT cgm" openaps.ini
katie disimone
@Kdisimone
May 09 2017 20:39
@diabetic_dude_twitter scroll up to this morning and see the nuclear option discussion. I posted the commands to remove them. BUT, you could just follow the docs for switching to the dev branch, updating it, and then just run the setup-script again. (I found that simply editing the runagain script didn’t bring in the new option flags for CGM…so running the setup-script was a good fix for that)
for switching to dev branch and updating it ^^
alternateal
@alternateal
May 09 2017 20:40
Having got SMB to work successfully using G5, I have been trying with MDT - I can't seem to get past this point - any thoughts?:
Starting supermicrobolus pump-loop at Tue May 9 21:39:43 BST 2017 with 14 second wait_for_silence:
Waiting up to 4 minutes for new BG: jq: error: null and number cannot be divided
date: invalid date ‘@’
glucose.json newer than pump_loop_completed
Radio ok. Listening: .
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 20:41
thanks I like docs and thanks for updating them @Kdisimone
katie disimone
@Kdisimone
May 09 2017 20:41
i like the docs too. :+1:
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 20:46
@Kdisimone like I said I updated the name on my NS site, I changed all of my IFTTT URLS, and will update the information in the setup script when it runs, is there anywhere else that I need to update my NS url information
katie disimone
@Kdisimone
May 09 2017 20:48
If you use xdrip…there. NS app on iPhone…there. That should be it, as far as I can think of.
rsilvers129
@rsilvers129
May 09 2017 20:54
Anyone know if Bluetooth on a PI3 with raspian is the same as on an Edison with Jubilinix? I ask because I want to know if people can do the development on a Pi and have the same code work for people with Edison.
David
@davidkeddydb
May 09 2017 20:58
I believe if you are trying to update your NS site and you use this: https://github.com/yourgithubname/cgm-remote-monitor/compare/dev...nightscout:floating-carbs (change dev to whatever your current cgm-remote-monitor/compare is and I think it will do it. I just followed the rest from the OpenAPS site and replace yourgithub name. Click the big green Create pull request button. Another screen will appear, fill in a title and click button to create the pull request, and then you can Merge pull request, and finally Confirm merge. That process updates your Nightscout repository in Github. Once updated, you can skip the “click the Fork Button” step below and start following along with the purple Deploy to Heroku button step from your updated Nightscout cgm-remote-monitor repository. Hopefully all of that is correct, but I think it is
Eric
@ecc1
May 09 2017 21:02
@rsilvers129 I set up a rig on a Pi Zero W recently, and used both BLE (to Dexcom G4 receiver) and BT tethering to my phone in the same way as the Edison (although not with the actual openaps code). The main issue you may run into is more limited RAM on the Pi, and if you have to swap to/from your SD card, it's pretty much game over.
rsilvers129
@rsilvers129
May 09 2017 21:10
I think p3 and Edison are both 1 GB ram.
Paul Dickens
@thebookins
May 09 2017 21:23
Hi All, can anyone give me some pointers as to how to read the BLE packets passing between the G5 transmitter and the Dex iOS app? Adafruit sells a "Bluefruit LE Sniffer", but I'm not sure if I need special hardware? I don't suppose there's an iPhone app that can do the same job (similar to hcidump on linux)?
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 21:24
@davidkeddydb what version did you end up with? and thanks BTW (I want to confirm I did mine right)
Eric
@ecc1
May 09 2017 21:24
@rsilvers129 no, only 512M
(for Pi Zero and Zero W that is)
You're right, should be OK on a Pi 3
katie disimone
@Kdisimone
May 09 2017 21:31
111__14_↑.jpg
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 21:33
0.9.3-dev-20161224 close enough right, kind of like having extra screws left over when your repair a laptop right?
Kate Farnsworth
@ELUTE
May 09 2017 21:35
not close enough lol
Diabetic_Dude
@diabetic_dude_twitter
May 09 2017 21:36
NOOOOO ok back to the drawing board
Tim Street
@tim2000s
May 09 2017 22:05
Looks like floating carbs has been pulled into the dev branch of NS.
Scott Leibrand
@scottleibrand
May 09 2017 22:13
@tim2000s where do you see that?
Martin Haeberli
@mhaeberli
May 09 2017 22:13
@thebookins FWIW, see also the ps2/rileylink channel, and the LoopKit/Loop channel, much less the related code. Because Loop can ask the Share App on the phone to share values as they come in… as I recall @ps2 knows all
Scott Leibrand
@scottleibrand
May 09 2017 22:14
@alternateal there is an oref0 PR open for fixing SMB to work with MDT. it's not yet working for @ericfeibelman - perhaps you could help debug
Tim Street
@tim2000s
May 09 2017 22:20
@scottleibrand my mistake. Looks like there are about 5 PRs for it into dev and I'd commented on one of them.
katie disimone
@Kdisimone
May 09 2017 22:21
@tim2000s lots of people arent doing their PRs right and sending in wrong ones.
Tim Street
@tim2000s
May 09 2017 22:21
@alternateal - that's usually an indicator that whatever your glucose.json from your SGV source is, it's containing null.
When I ran SMB with MDT SGV I manually amended the pump loop to pull in the SGV data as part of the pump history process. I haven't looked at this version.
fidoman
@efidoman
May 09 2017 23:17

@scottleibrand Hi Scott ... here's the results of that grep

root@eddie1:~/myopenaps# grep "MDT cgm" openaps.ini

get-bg = ! bash -c " (echo -n MDT cgm data retrieve && openaps monitor-cgm 2>/dev/null >/dev/null && grep -q glucose cgm/cgm-glucose.json && echo d) && cp -pu cgm/cgm-glucose.json cgm/glucose.json && cp -pu cgm/glucose.json monitor/glucose-unzoned.json && (echo -n MDT cgm data reformat && openaps report invoke monitor/glucose.json 2>/dev/null >/dev/null && echo ted)"

@scottleibrand - whoops wrong rig. my bad. the supermicrobolus one is eddie2. Anyway I think it is the same ...

root@eddie2:~/myopenaps# grep "MDT cgm" openaps.ini
get-bg = ! bash -c " (echo -n MDT cgm data retrieve && openaps monitor-cgm 2>/dev/null >/dev/null && grep -q glucose cgm/cgm-glucose.json && echo d) && cp -pu cgm/cgm-glucose.json cgm/glucose.json && cp -pu cgm/glucose.json monitor/glucose-unzoned.json && (echo -n MDT cgm data reformat && openaps report invoke monitor/glucose.json 2>/dev/null >/dev/null && echo ted)"
root@eddie2:~/myopenaps#

JELCRAWFORD
@JELCRAWFORD
May 09 2017 23:31
@Kdisimone - thanks for the pointer to that doc - is quite good - will bookmark and spend more time. Cheat sheet is great idea!
@scottleibrand @danamlewis - thanks for clarification on what that segment means - have seen Openaps decide basal rate is 'close enough' but were puzzled by the 2* bit. Cheers!