Hi,
First I apologize for this. The Atrium Gateway is still noted as Beta for reasons such as this. We are still solidifying over the air software updates.
Could you first ensure the sensor is still functioning? Remove the lid from the sensor and press the CFG button. You should see the LED inside the sensor turn on when the CFG button is pressed. This indicates it is on and the batteries are still good. If that works press and release the RST button. After a few seconds you should see that same STAT light in the sensor flash indicating it is sending data.
If you can confirm the sensor is in fact working by following the previous steps then open Node-Red on the gateway by opening your browser and entering ncd-a798.local:1880 Click the debug button which looks like a bug in the upper right corner of the window. Below that make sure the drop down is set to “all nodes”. Now press and release the RST button on the sensor. Let me know if you see any error messages in the debug log.
Thank you,
Travis
3/5/2026, 11:50:52 AMnode: Query devicesmsg : error
“Error: SQLITE_ERROR: no such column: d.last_status”
3/5/2026, 11:51:52 AMnode: Query devicesmsg : error
“Error: SQLITE_ERROR: no such column: d.last_status”
3/5/2026, 11:52:52 AMnode: Query devicesmsg : error
how do i factory reset the [Atrium-Gateway]
You will not want to factory reset the Gateway as that would clear all software.
It looks like the devices table schema for the SQLite DB table is missing the last_status column. This is why the device is not being entered into the table.
To fix this open the SQLite Console in the Atrium web interface. In the SQL Query field enter:
ALTER TABLE devices ADD COLUMN last_status TEXT;
Now clear that out and enter this in the SQL Query field to pre populate that column for all sensors with a default value:
UPDATE devices SET last_status = 'unknown' WHERE last_status IS NULL;
After doing this see if the Sensors page updates. If not check to see if there are any other issues in the Node-Red Debug log.
Let us know what you find.
Thank you,
Travis Elliott
unfortunately i tied to revert back to the existing version and now it looks like it’s back to 1.0.0 it’s not a big deal to reset we were using this unit in house for testing purposes.
Oh boy. I would say you’re going to need to send that Atrium Gateway back here at this point. The amount of time and frustration it would take to get this corrected at this point justifies returning the unit to us so we can correct it.
You can fill out an RMA here:
https://ncd.io/contact-us/product-returns/
Has there been any more development on this? I too did some updates, SQLite locked up all of my flows. I missed this thread, before resetting the Gateway and losing everything. I did backup all of my flows if I you guys have good way to get Node-Red reinstalled and whatever else I am missing. Any advice would be much appreciated. Thanks!
Do you still have access to the Atrium Dashboard?
If not you can re-flash Atrium onto the Gateway using a thumb drive.
The image and deployment steps are available here:
https://drive.google.com/drive/folders/1ZxcTeVU7EfxK1P-ESuIHnpHzlYcu4EdO?usp=sharing
If you have any questions on this process please let me know.
I am not having any luck re-flashing Atrium. I noticed that after the reset Node Red no longer works anymore. I tried to manually start it and this is what I’ve got. Any help would is greatly appreciated. Thanks!
In the terminal enter:
pm2 status
Please send screen shots of what you see from that.
Node-Red should be automatically started by PM2. You do not want to start in manually because it would conflict.
Yeah, definitely looks like there’s an issue Cloudflared has restarted 7,958 times and node-red has restarted 207 times.
Can you confirm that when you flashed from the thumb drive the LEDs on the front cycled in a circular pattern?
I think this may be due to not going through our validation sequence on new gateways. Your gateway may predate that validation sequence.
At this point I would recommend a manual OTA update via commands in SSH.
Before that though so I can collect as much information on what lead to this issue failure could you run this command in the SSH terminal and provide the response:
cat /overlay/.atrium-firstboot-done 2>/dev/null && echo "MARKER PRESENT (firstboot skipped)"
sqlite3 /overlay/telemetry.db "SELECT value FROM gateway_config WHERE key='platform_version';"
ls -la /middle/extend/home/ncdio/.pm2/dump.pm2 /home/ncdio/.pm2/dump.pm2
ls -d /home/ncdio/atrium-api /home/ncdio/atrium-ingest 2>/dev/null || echo "NO ATRIUM SOURCE ON DISK"
pm2 logs cloudflared --lines 20 --nostream
After that perform the manual OTA update:
cd /tmp && curl -fL https://github.com/ncd-io/Atrium-Gateway/releases/download/v2.3.2/gateway-update-v2.3.2.tar.gz -o gateway-update.tar.gz \
&& rm -rf gateway-update && mkdir -p gateway-update \
&& tar -xzf gateway-update.tar.gz -C gateway-update --strip-components=1 \
&& cd gateway-update && chmod +x apply-update.sh \
&& nohup ./apply-update.sh "$(cat version.txt)" > /tmp/gateway-update.log 2>&1 &
If the OTA update fails please run this command and provide the log returned
tail -f /tmp/gateway-update.log
So, after running the OTA Update, Node Red is accessible on 1880 and contains the flows for Atrium as well as the other stuff I was working on. The flows are stopped due to the SQLite error in the picture. This is similar to what happened after the SQLite nodes were updated in the palette. What would you advise as the next move? Also, Atrium does not appear to be available yet either. Thanks!
I just noticed something else odd. To do the OTA updates this was the first time that the gateway had been connected to the internet (since the other updates and reset) and that is when node red was accessible on 192.168.0.1, but when I disconnected the internet, node red loses connection to the server. Seems odd that an internet connection is required even though I am connected to eth1 and using the default 192 ip, I can understand this happening if I was using the 172.xxx.xxx.xxxx:1880 ip as I have had to do in the past for updates and other things, but just thought I’d let you know. Thanks again for working through all of this with me!
Here’s the most recent pm2 status.
If you open the gateway’s dashboard:
http://ncd-a783.local
Do you see Atrium’s Dashboard?
According to the logs Atrium-api and atrium-ingest are both running. These scripts completely replace the Node-Red flow as the backend of Atrium. You can actually turn Node-Red off and Atrium continues to function.
Do you intend to do something with Node-Red? If not I would recommend going into the Atrium Dashboard/Settings/Network and turning Node-Red off as it does consume more RAM and CPU than you might think.
Please let me know if Atrium Dashboard is up and running.
Thank you,
Travis
No the local ip does not work. I do use node red for more than just collecting ncd data. The flows are not able to run due to the missing SQLite nodes, see screenshot. I feel like with the current state of the flows Atrium is unable to restart. Let me know what you think. Thanks!
It appears the update stalled and did not complete.
I’m really not concerned with Node-Red or the flows not working at this point. Those flows would be removed upon a successful update. The issue is that the update is not completing.
When you power down the gateway, insert the thumb drive containing the eg5120-uuu-udisk-upgrade folder, then power the gateway back up what happens? After a few seconds the RUN LED on the front of the gateway should begin flashing rapidly(this indicates the udisk image is being flashed overwriting the entire file system) then after that completes the LEDs no the front of the gateway should begin flashing in a circular pattern. This entire process should take less than 2-3 minutes. Do you see any of this or does the Run LED just flash at a steady 1 second on 1 second off sequence continuously?
Please let me know.
Thank you,
Travis Elliott
After numerous attempts the gateway does not do anything differently on startup whether the thumb drive is inserted or not. It appears to do a normal startup that shows the run light solid for a few seconds then goes to the 1 second flash, none of the other indicators light up or show anykind of pattern.
Unfortunately we are getting mixed results on updating pre 2.x gateways to 2.3.2. Sometimes it requires inserting the thumb drive and power cycling a few times to get the fast flashing LED sequence that indicates the image flash is executing. If it’s not working I would recommend sending the Gateway back here so we can take care of this for you. You can fill out an RMA here:
https://ncd.io/contact-us/product-returns/
Thank you,
Travis