Open TheCranston opened 2 years ago
TheCranston
commented
2 years ago @rickkas7 Good morning. The uncaught exception restart of nodered is happening several times a day. Is there a deeper debug trace I can capture that will help? I wouldn't have known this is happening if I did have an event notification push to my phone on state change in the flow. Anyway, please let me know best how to proceed. thanks!
TheCranston
commented
2 years ago More findings. So it appears that the issue is around the re-authentication timeout. I found
res.body.expires_in: 7776000
which works out to be about 2 hours in ms. The same amount of time between faults and restarts. I over rode the value and set it to the max and it's been "holding" just fine. That's a work around and the core issue of a fault in the re-auth code persists. I'm not a javascript coder so I'm outta my depth on how to fix it or I'd have submitted a PR. Any and all help is appreciated. thanks!
TheCranston
commented
2 years ago Additional testing outside of homeasistant on a plain desktop install of node-red yielded the same results. However I was able to scope it down to the function nodes. the subscribe nodes do not exhibit this issue causing node-red to throw an uncaught exception. the desktop was running node-red v2.1.5 with node.js v17.3.0 on x64 linux
FranGarVal
commented
2 years ago I have the same problem when a migrate to v3 so I revert to 2.2.2 and is working again (node.js 16). I also tried another node from other author "node-red-contrib-particle", without success (similar problem?). Probably changed authentication particle's answer and newer versions are more unstable to the change.
TheCranston
commented
2 years ago @FranGarVal That has been my workaround as well. However it's not a long term option. Given https://cve.mitre.org/cgi-bin/cvekey.cgi?keyword=nodejs I'd like to keep reasonably updated to mitigate security issues.
FranGarVal
commented
2 years ago I installed in Raspberry Pi (Node v18.7.0) and exactly same problem :frowning:
Reauthenticating... Authenticated with Particle 20 Aug 15:21:23 - [red] Uncaught Exception: 20 Aug 15:21:23 - [error] TypeError: Cannot destructure property 'onEvent' of 'undefined' as it is undefined. at Api.listenToEventStream (/home/fran/.node-red/node_modules/@particle/node-red-contrib-particle-official/src/api.js:53:24) at login.then.retryIn (/home/fran/.node-red/node_modules/@particle/node-red-contrib-particle-official/src/api.js:169:29) at process.processTicksAndRejections (node:internal/process/task_queues:95:5) nodered.service: Main process exited, code=exited, status=1/FAILURE nodered.service: Failed with result 'exit-code'.
sberkovitz
commented
1 year ago I've fixed this in a fork. Until it is merged in, you can install direct with
npm install git+https://github.com/sberkovitz/node-red-contrib-particle-official.git
mjscoiot
commented
10 months ago More findings. So it appears that the issue is around the re-authentication timeout. I found
res.body.expires_in: 7776000which works out to be about 2 hours in ms. The same amount of time between faults and restarts. I over rode the value and set it to the max and it's been "holding" just fine. That's a work around and the core issue of a fault in the re-auth code persists. I'm not a javascript coder so I'm outta my depth on how to fix it or I'd have submitted a PR. Any and all help is appreciated. thanks!
How did you override this setting?
TheCranston
commented
10 months ago @mjscoiot in api.js I simply used the max value rather than lower of max and the expires_in value received. nasty hack. I'm not a java programmer for certain. A far better approach is what @sberkovitz did in his fork. I'd consider that the correct path as it actually addresses the exception. Let's hope that his PR is accepted.
mjscoiot
commented
10 months ago I ended up just commenting out everthing in the _reauthenticate function
after .login such that basically it doesn't retry the login within that
section of code in
@.***/node-red-contrib-particle-official/src/api.js`.
Seems to work flawlessly since login rarely fails.
On Sat, Dec 23, 2023 at 7:21 AM TheCranston @.***> wrote:
@mjscoiot https://github.com/mjscoiot in api.js I simply used the max value rather than lower of max and the expires_in value received. nasty hack. I'm not a java programmer for certain. A far better approach is what @sberkovitz https://github.com/sberkovitz did in his fork. I'd consider that the correct path as it actually addresses the exception. Let's hope that his PR is accepted.
— Reply to this email directly, view it on GitHub https://github.com/particle-iot/node-red-contrib-particle-official/issues/3#issuecomment-1868293714, or unsubscribe https://github.com/notifications/unsubscribe-auth/BE3367HLOBIAIRAWNLSRH33YK3LEBAVCNFSM5L4LWV7KU5DIOJSWCZC7NNSXTN2JONZXKZKDN5WW2ZLOOQ5TCOBWHAZDSMZXGE2A . You are receiving this because you were mentioned.Message ID: <particle-iot/node-red-contrib-particle-official/issues/3/1868293714@ github.com>
I've had the particle integration via node-red working for quite a while now and at some point a guess a version change happened and now every so many hours I get this cycle of "Reauthenticating" then "exception" and finally node red restarts.
Not quite sure what's happening. Below is the log of one of these events.