@scottleibrand i tried both the differences you have in your interfaces file. Same result. connects but not able to ping.
here is my wpa_supplicant
ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
update_config=1
network={
# Home in SF
ssid="jsqrd"
psk=<PASSWD>
priority=3
}
network={
ssid="James' iPhone"
psk=<PASSWD>
priority=2
}
network={
ssid="belkin.81c"
psk=<PASSWD>
priority=3
}
it works for the first two entries.
@billstackpole Standard markdown. Can use GUi web github for docs if easier.
i did try a bunch of different things in there from Google searches. The belkin is a “WPA/WPA-2” where as my other network is a WPA-2.
are your PASSWDs "in quotes", or just plain text?
plain text hashed (no quotes). it is able to connect and get on the network.
including the belkin one?
yes
at what step does it fail on that network?
the logs say it connects and the router’s connected devices lists my edison
let me find the logs.
but what doesn't work?
does it get an IP address?
can you ping the router?
i can not connect to the internet. ping google.com does not work
i can try to ping the router. i have not tried that before.
and the 192.168.x.y address you pasted earlier: what were you connected to then?
the only error i can see in the logs is
localhost avahi-daemon[1542]: No service file found in /etc/avahh
i/services.
i am always connected to the same ip when i reboot the edison and it comes back up
i can ping the router
are you saying the IP of the edison doesn't change when you switch networks?
no. i am saying that whenever i connect to this home wifi i always get the same IP. if i were to connect to the hotspot i would get a totally different IP. i am saying it seems to be working as expected.
ok, good. what IP is the belkin giving you when you connect to it?
192.168.2.59
and presumably the router itself is 192.168.2.1?
yup
try
traceroute google.com
if you can ping the router, then WPA is fine
there's something else going on
edison@mursefree:~$ traceroute google.com
google.com: Name or service not known
Cannot handle "host" cmdline arg `google.com' on position 1 (argc 1)
try
traceroute 8.8.8.8 then
your DNS resolution is failing
it’s going really slow. first four are all *’s then it started hitting some Ips
ok, if you want you can cancel that and try
ping 8.8.8.8
but it looks like you're getting out to the Internet, you just don't have valid DNS
traceroute to 8.8.8.8 (8.8.8.8), 30 hops max, 60 byte packets
1 * * *
2 * * *
3 * * *
4 * * *
5 * 96.34.121.156 (96.34.121.156) 62.718 ms 64.437 ms
6 96.34.121.50 (96.34.121.50) 64.364 ms 19.527 ms 22.308 ms
7 96.34.1.4 (96.34.1.4) 38.176 ms 29.525 ms 45.807 ms
8 96.34.3.1 (96.34.3.1) 31.111 ms 44.475 ms 30.488 ms
9 72.14.220.11 (72.14.220.11) 24.882 ms 31.058 ms 35.225 ms
10 209.85.244.23 (209.85.244.23) 33.321 ms 209.85.241.171 (209.85.241.171) 24.058 ms 26.295 ms
11 216.239.49.121 (216.239.49.121) 31.903 ms 216.239.58.193 (216.239.58.193) 38.752 ms 64.233.174.43 (64.233.174.43) 36.874 ms
12 8.8.8.8 (8.8.8.8) 28.657 ms 41.342 ms 41.384 ms
you could set some of the hosts you need in
/etc/hosts
if you can't get dns working
yeah, but shouldn't have to do that
follow the Debian server directions
Hi, all. I'm working my way through the install docs and have two questions about Nightscout integration: 1) am I correct to assume that the NS source is installed on the RPi along with OpenAPS? 2) does the profile editor bug still exist in the dev version of NS? Looks like that warning was added in late January, so I'm hoping the problem has been resolved.
No, NS is intended to run in the cloud, not locally on the Pi.
Oh, duh. Now the Azure and Heroku stuff makes sense.... Thanks.
What does it mean if
ifconfig, ifdown wlan0, or ifup wlan0give command not foundin the Edison?
Your PATH is messed up.
You can
sudo bash first. Or go fix /etc/profile
it defaults to a stupid PATH with stuff like games, not system directories.
We should probably add something to the docs to fix that.
@dougtakeuti I think all of those are in /sbin
You need to be
sudod to run ifup and ifdown and on my Edison, when you sudo it finds the commands.
But I also added
/sbin to my path to avoind annoyance.
Thanks, @scottleibrand i can ping google now after editing the /etc/dhcp/dhclient.conf file per the link you sent
some of the info in there did not work but that part seems to work and stick after a reboot
next on the list is to try to get ssh working. i can’t ssh to the .local address (or directly to the IP)
i can ping it, but can’t ssh
thinking it may be the firewall
I can't find any docs on it, but what's the update on G5 receiver working via a wired connection?
I remember reading somewhere that they just needed to put in the autodetect but it was possible to hardcode it?
NS works.. but it's a bit spotty here and there
working now. reset my connection on the laptop and must have found it i guess.
now to get the TI stick going on my intel board
i can’t seem to get the TI stick recognized. No lights and i dont think it is connected on either USB port (I know only one of them should work).
guess i am in the wrong thread for this question :)
heh
@itsn0tm3 G5 state is still the same. There's a dexcom_reader branch you can use, if you only use G5. No autodetection yet, so not merged.
Oooh I only use G5 so that might work
where is the branch at?
Don't recall. Try a gitter search for it.
will do, thanks for the hint!
Hi, this is @logichammer 's wife, and we have a question about getting the nightscout dev branch working on azure. Once we have a profile established on master, how do we go about switching to dev? Do we need to disconnect the master deployment, and then do a new deployment with dev? Any help would be appreciated!
yes, if you go to settings -> continuous deployment (under publishing) -> then in upper right press disconnect. then you can connect it to dev
Thanks for confirming that. I did go through those steps and am getting a white screen on the nightscout site for some reason. It was working when setup on master so I know the mongo string is correct. Azure says the deployment was successful, too.
Does Chrome Dev Tools show anything informative on console?
It says 'Failed to load resource: the server responded with a status of 500 (Internal Server Error)'
are you sure your dev branch is current and error free?
Make sure you've merged the latest dev to your dev. Merging master doesn't update it.
And if that is good, what URL is it getting a 500 on?
We did merge the latest dev first. Does it matter if the master I started with is on the latest version? Ours was 15 commits behind.
I find it easiest to delete your fork and then refork the source
That way you don't have to worry about merging
Your fork doesn't have any commits that aren't merge commits, so that should be safe.
But I confirms your dev branch is up to date,
so that may not help.
*confirmed
May be a mongo thing? Dunno. The CGM in the Cloud support folks are better for this.
Thanks @eyim , we'll try starting a new fork and see if that works. Matt says hi.
And thanks for the help @scottleibrand !
The URL of that 500 error in dev tools might also be a clue.
Hi All. I've got a Pi 3 and a Pi Zero now, and a 722 plus carelink stick at the post office waiting for me to collect. The initial plan is to set up the Pi 3 plus carelink, following the official docs closely (the only difference is I'll be pulling CGM data from Nightscout and I've got that working now on the Pi 3).
The Pi Zero will be my portable rig, but I'm not sure if I should go for Slice of Radio, ERF/SRF or TI stick? I'm okay with some basic soldering and I'm looking for something that's quite compact
Slice of radio should work with the Pi Zero, but I'm not sure if anyone else has tried that?
I suppose the main consideration is flashing the firmware as I don't currently have access to the tools to do that. Are there any options which I wouldn't need a good FET or similar for?
I should maybe ask this question over on mmeowlink too!
@jasoncalabrese when I upload to heroku, I see two add-ons: the usual mLab and the Papertrail log service. Do we have a similar log system/service when uploading to azure? I don't see any collections in mLab that seem to relate to that, so I assume the answer is no.
I reached my 10MB daily Limit pretty quickly and it appears to me that the reason is this:
https://gist.github.com/dougtakeuti/3981a368f280e25dd77e
Apparently I still have formatting issues with the glucose.json files that I upload. What do I need to look for to fix that problem?
I reached my 10MB daily Limit pretty quickly and it appears to me that the reason is this:
https://gist.github.com/dougtakeuti/3981a368f280e25dd77e
Apparently I still have formatting issues with the glucose.json files that I upload. What do I need to look for to fix that problem?
the 10MB limit is paper trail?
everyone is getting that "found a glucose value without any units" message since the upload from openaps doesn't include units
maybe we can hardcode it to mgdl, but I haven't done it since there have been so many issues with units in the past and expect a problem any time something with unit is changed
think we just need to add unit here https://github.com/openaps/oref0/blob/dev/bin/mm-format-ns-treatments.sh#L18
@cpressnall you might have to update your mongo string. Mongolab changed their name to mlab. Updating your site may have triggered the name change.
@jasoncalabrese yes, the 10 MB is for papertrail. Is that also the limit for Heroku? Will it run without the log?
will treatments always be in mgdl, I'm not sure how it looks with mmol, I'd like to fix it
I thought there were several ways people were handling mmol
should just always run the units converters
what's the result of that? all BGs converted? will always be sent to NS as mgdl?
hmm, don't think the
units device works on pump history, but for bg-targets, sensitivity and others, it always converts the units
in those cases medtronic is always very explicit about what units the data are in
so it's safe to always convert, (the converter is a no-op in the case where selected units is already used)
if we send the units to NS it will use it
nice to avoid the conversion where possible
I think that might be what's happening, it's sent without units, so then uses server default
yeah ,so just should ust annotate using pump-settings.json or similar
the units being used is described there, so it could be used to annotate all the pump-history
My Carelink USB keeps on failing, even though I running
sudo oref0-reset-usb in my preflight alias. When I look through the cron log, I see it happening but it never fixes itself, it just keeps on saying that the carelink stick is failing.
it's plugged directly into a PI?
And is it in range?
@jasoncalabrese It is plugged into the pi through a usb dongle, does that change things?
@danamlewis Yeah, it happens when I'm in range generally.
@danamlewis Yeah, it happens when I'm in range generally.
try a new USB cable maybe
I was just wondering if it was on a hub, since you'd need to shut power to the hub
saw that when I was trying to use a pi zero
I don't think its the USB cord, but I'll try another one just to be sure. My rig will work consistently for hours, and then the carelink will go dead until I replug it in again, even though I'm running the reset USB command.
running the command manually doesn't work either?
No
sounds like power isn't getting cut off from the stick for some reason
Although when I'm looking through the cron logs, it says this:
dexie://JSON/iter_glucose/monitor/glucose.json
reporting monitor/glucose.json
pumpy://JSON/read_clock/monitor/clock.json
/usr/local/bin/mm-stick: line 119: /dev/fd/1: Permission denied
Power-cycling USB to fix dead Carelink stick
PREFLIGHT SLEEP
PREFLIGHT FAIL
No CGM data.Is the permission denied there normally?
nope. you need to sudo that
Hmm, my preflight alias shows that I am though:
pi@raspberrypi:~/apsfallon $ openaps alias show preflight
preflight ! bash -c 'rm -f monitor/clock.json && openaps report invoke monitor/clock.json 2>/dev/null && grep -q T monitor/clock.json && echo PREFLIGHT OK || (mm-stick warmup || (sudo oref0-reset-usb && echo PREFLIGHT SLEEP && sleep 120); echo PREFLIGHT FAIL; exit 1)'
That's weird. And when I run it from the command line, I don't get that error, but it still doesn't work. Would rebooting the Pi be a viable solution for this, instead of running the reset usb command?
maybe sudoers isn't right
the preflight from my old pi looks the same
you're not doing sudo on mm-stick, which is what is erroring
shouldn't have to though
looks like maybe device perms aren't right?
I didn't have a sudo for mm-stick only the reset
when you sudo after first logging in does it ask for your password?
No
And it looks like sudoers is just the default file
try
mm-stick warmup interactively?
It just pauses for a second and does nothing
got a question - just turned on my Pi and trying to run gather-profile - getting a "attempting to use a port that's not open" message
thinking this might be related to low power, usb on an extension cable, Carelink weirdness or otherwise - anyone else have experience?
probably multiple commands using the radio at the same time