Netonix Manager / Control Software

[tomnbb]

Are you executing those statements as root?

That's why it says [root@xms netonix-manager]#

[Stephen]

Running npm update may have caused incompatibilities in the software if any of the library function's in the node_modules where deprecated.
Try overwriting the netonix-manager with the tar.gz again and then run the commands above. Let's see how that goes.

[tomnbb]

Code: Select all

[root@xms netonix-manager]# curl -O https://forum.netonix.com/netonix-manager/netonix-manager-1.0.14.tar.gz
[root@xms netonix-manager]# tar -zxvf netonix-manager-1.0.14.tar.gz --no-same-owner -C /
[root@xms netonix-manager]# export NODE_ENV=production
[root@xms netonix-manager]# node bootloader.js
starting child process
{ Error: Cannot find module '../build/Release/bson'
 at Function.Module._resolveFilename (module.js:548:15)
 at Function.Module._load (module.js:475:25)
 at Module.require (module.js:597:17)
 at require (internal/module.js:11:18)
 at Object.<anonymous> (/opt/netonix-manager/node_modules/bson/ext/index.js:15:10)
 at Module._compile (module.js:653:30)
 at Object.Module._extensions..js (module.js:664:10)
 at Module.load (module.js:566:32)
 at tryModuleLoad (module.js:506:12)
 at Function.Module._load (module.js:498:3) code: 'MODULE_NOT_FOUND' }
Express server listening on port 3443

I can login and it works, the main reason I had done the updates before was the bson issue. I think you can use packagelock.json to force it to a certain version if updates break you, though. Doing npm install bson to just get that one breaks it, which version of bson are you trying to use? I think I can install @ that version.

Code: Select all

starting child process
{ Error: Cannot find module '../build/Release/bson'
 at Function.Module._resolveFilename (module.js:548:15)
 at Function.Module._load (module.js:475:25)
 at Module.require (module.js:597:17)
 at require (internal/module.js:11:18)
 at Object.<anonymous> (/opt/netonix-manager/node_modules/mongoose/node_modules/mongodb/node_modules/bson/ext/index.js:15:10)
 at Module._compile (module.js:653:30)
 at Object.Module._extensions..js (module.js:664:10)
 at Module.load (module.js:566:32)
 at tryModuleLoad (module.js:506:12)
 at Function.Module._load (module.js:498:3) code: 'MODULE_NOT_FOUND' }

##############################################################
#
# !!! MONGOOSE WARNING !!!
#
# This is an UNSTABLE release of Mongoose.
# Unstable releases are available for preview/testing only.
# DO NOT run this in production.
#
##############################################################

process exit code 1

Pro tip, you might want to run some checks in start.sh to see if node is nodejs or node on this env, and possibly check the version but that could be overkill. I had trouble and found that it was using nodejs instead of node.

[Stephen]

Netonix Manager version 1.0.15 is now out.

We skipped a few versions due to some internal testing. Like normal, check the release notes for details.

There was an issue we found in the NM that was capable of causing a switch to lock up randomly, so please update when you can to prevent this from occurring.

Also, there was an issue with the updating procedure that had to be repaired.
The way the fix works, you should be able to perform an update now by doing it twice from whatever version you're on.
(Click on the question mark icon on the top right, it should tell you that a new update is available, go through the procedure and then when it restarts do it once more and then the NM should be at version 1.0.15)

[SeanDB]

Would it be possible to add the voltage to the manager main screen? Would be handy to see at a glance if solar sites are looking sad...

[Stephen]

I will investigate the option. Added to the list.

[yahel]

Is it normal for Netonix manager to generate high traffic (traffic returning from the switches to it)?

The attached graph shows ONLY Netonix-Manager traffic --- shutting down the app stops this traffic completely.
This particular one manages about 20 switches...

[Stephen]

If you're Netonix Manager is watching the status of all of those switches then it has to continually request a json file from each of them that contains the status details of that switch. If this is too cumbersome on your network. You can reduce the impact editing a few of the switches and turning off the "Monitor" option. Which will prevent the NM from continuously polling them but won't prevent you from doing upgrade's, accessing the web UI etc.

[yahel]

Na - this started recently... something is wrong...
Also - seriously, a peak of 3mbps for a second or so, means a VERY large json file even if divided by 20....

Assuming you are correct, that would be devastating in my other network where NM is monitoring ~100 switches, and it does no such thing.
It would also suggest a pretty poor design -- why pull everything every second, where you should only pull diffs (rsync like) -- I'm fairly certain Eric would not have released such functionality... You must be guessing... or if indeed it work this way, the json must be so tiny (around a single packet or so), and not create such traffic under normal conditions...

[Stephen]

Well, let's check something, about how large are the log file's on the switches being monitored? This is something we've been watching. It's possible depending on what's happening the NM might be trying to pull in the log file and of course, over time these progressively grow. Are the log files on the switches very large? And if they are, if you clear them - does that stop these spikes?