Repository navigation
Replies: 5 comments 15 replies
|
Good find. Fixed. Pull the code. |
|
Thank you for your detailed instructions and explanation. Unfortunately, something is still amiss and I do not have a clue how to resolve it. Below is a screen scrape with the commands you posted above and the system responses. Step 4 resulted in errors so I jumped to step 7. It also resulted in the same results. I am truly trying very hard not to be difficult am I'm willing once again to flash a new OS copy & start over, but I'm missing something along the way or out and out doing something incorrectly. Unfortunately, I don't know what it may be. I have been following the Wiki as closely as I can to get a fresh system going but for whatever reason, it's not working for me. As always, I am very grateful for your patience and help! michael@nixie:~ $ cd added 431 packages, and audited 432 packages in 3m 96 packages are looking for funding found 0 vulnerabilities michael@nixie:~/relayEquipmentManager $ node -e "require('ip'); require('node-ssdp'); console.log('ip + node-ssdp OK')" Error: Cannot find module 'ip'
Node.js v20.20.1 added 431 packages, and audited 432 packages in 3m 96 packages are looking for funding found 0 vulnerabilities Error: Cannot find module 'ip'
Node.js v20.20.1 |
|
I don't know why there are strikeouts in some of the lines above. Please ignore them as the commands/responses are actual cut & paste from the terminal window and none of the text has strikeouts. |
With help from AI.Yep — that response is the smoking gun. From the user’s own output in that comment, REM is installing node-ssdp@1.0.0, not 4.0.1 (discussion comment).
REM Dependency Mismatch Fix (node-ssdp should be v4, not v1)If 1) Confirm what files your current folder is actually usingcd ~/relayEquipmentManager
pwd
git remote -v
git branch --show-current
git rev-parse --short HEAD
node -p "require('./package.json').dependencies['node-ssdp']"
node -p "require('./package-lock.json').packages['node_modules/node-ssdp']?.version"Expected values:
2) If either value is wrong, force-refresh from origin/mastergit fetch origin
git checkout master
git reset --hard origin/masterIf you want to avoid any reset risk, do a fresh clone instead: cd ~
mv relayEquipmentManager relayEquipmentManager.bak.$(date +%s)
git clone https://github.com/rstrouse/relayEquipmentManager.git
cd relayEquipmentManager3) Confirm npm registry is correctnpm config get registryExpected:
If not: npm config set registry https://registry.npmjs.org/4) Clean install from lockfile onlyrm -rf node_modules dist
npm cache verify
npm ci5) Verify installed versions and runtime resolutionnpm ls node-ssdp ip
node -p "require('node-ssdp/package.json').version"
node -e "require('ip'); require('node-ssdp'); console.log('ip + node-ssdp OK')"Expected:
6) Build and run REMnpm run build
npm start |
|
OK, yep it's early in the AM, even here. OK, I think I figured it out! When I started the processes using npm start, it responded with various errors, so I ran npm audit fix. More errors, so I ran npm audit fix --force. Bad idea!!!! As below, if you force the audit fixes it does the following: Severity: high And that is what messed up the world for me. Killed everything again but I only fixed audit errors that didn't require the --force switch. And the system builds & is running (getting it back slowly, anyway). It may be worth putting a note in the wiki about this. Oh, what a world! Thanks for your patience & help! |
|
Ya, generally running |
|
Thanks for your quick response! Haven't been able to get things running fully yet because NPM isn't happy & the upgrade attempts for NPM failed. Also having issues with temperatures not updating in DashPanel, REM values seem correct. May try a fresh build from scratch to get everything going again. Sigh. |
|
Thanks for your help!! About to dive back in & hoping to report success
soon!
…On 3/15/26 9:46 PM, tagyoureit wrote:
Drop back to Node v20. Haven't tested 22 or 24 yet. Then completely
kill the node_modules directory and run npm i again.
—
Reply to this email directly, view it on GitHub
<#1159 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AJ3NIBD6643L6NUVPWD235T4Q5THVAVCNFSM6AAAAACWP62ENOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTMMJVGA2DGMY>.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
Thx! I will give it a whirl in the AM. BTW, which AI are you using?
Chatgpt has its strong points, but is not so great with technical
challenges. Ask me how I know 😖
Hope to report good news to you tomorrow!
…On Mon, Mar 16, 2026, 8:49 PM tagyoureit ***@***.***> wrote:
With help from AI.
Yep — that response is the smoking gun.
From the user’s own output in that comment, REM is installing
***@***.****, not 4.0.1 (discussion comment
<#1159 (reply in thread)>
).
- ***@***.*** depends on ip (expected in current REM lockfile)
- ***@***.*** does not include ip
- so require('ip') fails exactly as shown
REM Dependency Mismatch Fix (node-ssdp should be v4, not v1)
If npm ci is giving ***@***.***, your local source-of-truth is wrong
(wrong files, wrong branch/repo, or wrong working directory).
Follow this in order.
1) Confirm what files your current folder is actually using
cd ~/relayEquipmentManagerpwd
git remote -v
git branch --show-current
git rev-parse --short HEAD
node -p "require('./package.json').dependencies['node-ssdp']"
node -p "require('./package-lock.json').packages['node_modules/node-ssdp']?.version"
Expected values:
- package.json dependency: ^4.0.1
- package-lock.json node_modules/node-ssdp version: 4.0.1
2) If either value is wrong, force-refresh from origin/master
git fetch origin
git checkout master
git reset --hard origin/master
If you want to avoid any reset risk, do a fresh clone instead:
cd ~
mv relayEquipmentManager relayEquipmentManager.bak.$(date +%s)
git clone https://github.com/rstrouse/relayEquipmentManager.gitcd relayEquipmentManager
3) Confirm npm registry is correct
npm config get registry
Expected:
- https://registry.npmjs.org/
If not:
npm config set registry https://registry.npmjs.org/
4) Clean install from lockfile only
rm -rf node_modules dist
npm cache verify
npm ci
5) Verify installed versions and runtime resolution
npm ls node-ssdp ip
node -p "require('node-ssdp/package.json').version"
node -e "require('ip'); require('node-ssdp'); console.log('ip + node-ssdp OK')"
Expected:
- ***@***.***
- ip present
- console prints ip + node-ssdp OK
6) Build and run REM
npm run build
npm start
—
Reply to this email directly, view it on GitHub
<#1159 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AJ3NIBFLEYF3OKKN4OSSO534RCVJFAVCNFSM6AAAAACWP62ENOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTMMJWHE4DCNQ>
.
You are receiving this because you authored the thread.Message ID:
<tagyoureit/nodejs-poolController/repo-discussions/1159/comments/16169816@
github.com>
|
|
I typically use Cursor with either the Claude Sonnet 4.6 or GPT 5.3-codex models. I've got a bunch of rules/skills on the project so all of the questions I ask the LLM come pre-loaded with context and other critical information about the project. |
|
OK, I decided to try a slightly different path to success. For the past few days, I've been installing a 64 bit version of raspian on my Pi 3b+. That's different than my previous (working!) configuration. Some chatbot convinced me things would be better if I switched to 64 bits. When I built a Nixie for the first time 2-3 yearas ago, I believe @rstrouse suggested that a 64 bit OS was less stable. After my past few days, I'm beginning to think that's still a true statement. So, back to 32 bit raspian I went, starting from scratch once again. I skipped the npm audit fix --force repairs & everything appeared to start OK. Once again, however, the I2C bus in REM displays all possible addreseses as device candidates (rather than all/some of the actual boards that are attached). I configured the megabas inputs and fed them to the temperature probes on the 'generic' tab. However, they show temperatures of 0 and the inputs on the megabas show 0 ohms. Odd, because in the 64 bit OS I was getting reasonble temperatures. Similarly with the flow switch. And now, the relayEquipmentManager is throwing errors, pooping out, restarting, & repeating the cycle. The only times I've touched the Pi stack were to remove/insert the microSD card. I've examined the stack & all the header pins seem to be aligned correctly & none appear to be bent (again, it's been a week or more since I restacked the boards & I received good inputs even when the system pooped out). No wires in the controller box appear to be detached or damaged. Here is the output from REM when starting the process via npm (similar to using PM2, but it's more readable since the process doesn't restart. michael@nixie:~/relayEquipmentManager $ npm start
linux Do I need to re-stack the boards? The GPIO pin messages concern me. Anything else I can do? Thanks once again for any/all thoughts! Hoping for a happy controller soon. |
|
Wednesday noon update: I was planning to take the stack apart & run "1c2detect -y 1" on just the pi to see if I could locate the problem with all of the addressses showing activity, if all was good with just the pi, I would add the HATs one at a time & repeat the process until I found whatever was misbehaving. However, when I powered on the stack a few minutes ago, the results were correct, as below: michael@nixie:~ $ sudo i2cdetect -y 1 I don't know why (meteors hitting Cleveland fixed the problem, perhaps?). Anyway, I ran "PM2 start all" & then "pm2 monit" to see what's happening & working. njpPC and dashPanel are up & running without any issues and the pool controller shows up at ipaddress:5150. REM, however, is still not happy. The logs are below. In short, REM appears to start, then runs into problem, crashes, restarts, & the doom loop continues until I kill the process. REM > > relay-equipment-manager@8.3.0 start │ REM, like njsPC and dashpanel, were freshly optained & built yesterday from the git repositories. I believe I'm running the correct versions of npm and node.js, as below: michael@nixie:~ $ node -v So where do I go from here to get things operating correctly? Thank you for your patience & help!! |
|
Everything started fine with npm start but when I configured pm2 &
started all, REM keeps spitting out errors & restarting. Stuck again
at the moment.
REM also showed all I2C numbers as unknown devices. I configured
(partially) the ones I have--megabas, relay board, temperature probes,
flow switch. Can't get 485 communication going either btn REM &
poolController. Will try editing poolController config.json to point to
correct REM IP address there (pi is not called 'raspberry').
Suggestions continue to be appreciated. I don't recall things being
quite this challenging when I first build Nixie :-(
From pm2 monit:
REM > [17/03/2026, 10:47:33] info: Server is now │
│[ 1] dashPanel ││ REM > Error: EINVAL: invalid argument, write
│
│[ 2] njsPC Mem: ││ REM > at Object.writeFileSync
(node:fs:2380:20) │
│[ 3] pm2-logrotate ││ REM > at exportGpio
(/home/michael/relayEquipme │
│ ││ REM > at new Gpio
(/home/michael/relayEquipment │
│ ││ REM > at GpioController.initPin
(/home/michael/ │
│ ││ REM > at GpioController.initPins
(/home/michael │
│ ││ REM > at GpioController.init
(/home/michael/rel │
│ ││ REM > at
/home/michael/relayEquipmentManager/ap │
│ ││ REM > at processTicksAndRejections
(node:intern │
│ ││ REM > > ***@***.*** start
…On 3/17/26 10:37 AM, tagyoureit wrote:
Ya, generally running |npm audit fix| will install things that you
don't want.
—
Reply to this email directly, view it on GitHub
<#1159 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AJ3NIBCBMQXEFRPTRZ27BQT4RFWM5AVCNFSM6AAAAACWP62ENOVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTMMJYGAYDOMQ>.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
Did you follow the PM2 wiki? |
|
I did. Oddly, I can start things about 1/2 of the time without errors. otherwise, I get the error above. Very odd--not stopping anything at the moment, because some stuff seems to be working, but not all :-( See my what the pup thread for more tales of woe. |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Finished migrating Nixie & REM to a relacement Pi, however the status of the acid pump relay is not updated when the relay turns on. I can turn the pump on from either the DashPanel or the multi-relay board manually, and it does turn on automatically when it needs to dose acid. However, when Nixie/REMChem doses acid, the acid feeder dot never turns green to show the pump is on. All other relay statuses are accurate on Dashpanel. I'm probably missing something obvious, but I've not figured out what. Please help! Many thanks!!! Screenshots of configuration settings attached.
P.S. Interestingly (mabe?) Nixie shows that REM is dosing acid accurately, even though the relay state is not correctly displayed.
P.P.S. I don't have a screenshot, as expected in the Chemistry section of DashPanel, the REM Dot does turn green when the acid pump is running. Why it is correct and the Features section showing the Acid Pump never turns green I don't know.
All reactions