cmon bitdefender

It was definitely unrelated to the netclient service stopping on several nodes, which was induced by Cogent maintenance causing packet loss on my VPS

Going to test stopping the broker later to induce the issue and try setting a recovery method for the windows service

Windows is a pain in the bum..

On the nodes affected, the service was stopped following the events shown in my pastebin log above

ran a 'docker-compose stop mq' test and started it after 15-20 seconds

the netclient service on every windows node stopped and never restarted, even with recovery methods set

all nodes are win10/11, v0.14.4

something about the netclient daemon or winsw causes this

testing different configuration for it now.. also trying nssm as a service manager

okay this might have something to do with it

notice once the daemon tries to re-establish, it tries to reach the broker on 8883 instead of 443

currently using the traefik compose

opening port 8883 on server fixed it, but i assume this wasn't the intention

mq port set to 443

humorously bitdefender has become very dumb and refuses to acknowledge an exception for netclient

this is how i know your software is especially malicious feiszli

🤔

positively false false positive

okay that did not help the service problem. if it loses contact with broker, the log shows

[netclient.exe] 2022-07-02 01:28:39 error with stop of Windows Netclient daemon: exit status 0xc000013a : 2022-07-02 01:28:39,109 INFO  - Stopping service 'netclient (netclient)'...
2022-07-02 01:28:39,118 INFO  - Service 'netclient (netclient)' stopped successfully.
^C 
[netclient.exe] 2022-07-02 01:28:39 running start of Windows Netclient daemon 
[netclient.exe] 2022-07-02 01:28:39 successfully ran start of Windows Netclient daemon

but netclient did not start

linux nodes re-establish beautifully. windows is necessary evil

managed to get a nssm created service to keep contacting the broker

worked fine after stopping mq and waiting.. tried it several times

stdout shows netclient daemon trying to start the winsw service again repeatedly and fails, but that's fine in nssm's service keeps the process alive

well for anyone having a similar issue and who finds this NSSM (https://nssm.cc) is my solution for now.. annoyingly it defaults to opening a GUI and i have yet to learn its command syntax but it works

shutdown mq for a good 5+ minutes then started it again and every node found its way back to healthy status

oh and set the service to automatic delayed start otherwise wireguard has issues

maybe starts to too fast.. log shows this:

[netclient.exe] 2022-07-03 01:33:07 failed to parse interface 
[netclient.exe] 2022-07-03 01:33:07 error encountered checking local listen port:  The program issued a command but the command length is incorrect. 
[netclient.exe] 2022-07-03 01:33:07 failed to parse interface 
[netclient.exe] 2022-07-03 01:33:07 error syncing wg after peer update: The program issued a command but the command length is incorrect. 
[netclient.exe] 2022-07-03 01:33:07 failed to parse interface 
[netclient.exe] 2022-07-03 01:33:07 error syncing wg after peer update: The program issued a command but the command length is incorrect. 
[netclient.exe] 2022-07-03 01:33:22 failed to parse interface 
[netclient.exe] 2022-07-03 01:33:22 error syncing wg after peer update: file does not exist 
[netclient.exe] 2022-07-03 01:33:59 failed to parse interface 
[netclient.exe] 2022-07-03 01:33:59 error encountered checking local listen port:  file does not exist 
[netclient.exe] 2022-07-03 01:33:59 checkin for stbe complete 
[netclient.exe] 2022-07-03 01:34:22 failed to parse interface 
[netclient.exe] 2022-07-03 01:34:22 error syncing wg after peer update: file does not exist

I'll give it a try and let you know