I suspect that it is in the grpc code, as my GUI (https://github.com/mattkasun/netmaker-gui) which includes alot of the netmaker code but not the grpc stuff, does not suffer from the issue

The GUI doesn’t hit any grpc endpoints so not sure if that would explain it

We’re aware of the memory issue and are working on a refactor right now which will take care of it. I think this issue is not in 0.8.5 so that may be a better option but I’m also thinking there’s some simple docker stuff we can do to restart the container once it hits X memory. The container starts really fast so shouldn’t cause any noticeable issues and will be a lot quicker of a fix

we can pretty easily set a memory limit in docker-compose, though not sure if hitting the limit will trigger an automatic restart: https://www.baeldung.com/ops/docker-memory-limit

My point was that my GUI does not have the memory leak so the leak must be in that part of netmaker which my GUI does not access, that being the grpc code

so netmaker seems to idle around 46m, I added mem_limit: 70m specifically to the netmaker container, we'll see if that catches it

sidebar.... what's the deal with the dualing netmaker nodes? For every network I create, I seem to end up with two netmaker nodes

the initial x.x.x.1 node is always in error, and netmaker seems to create another node that ends up as the last IP in the network address range

that's weird, I've never seen that one...what version are you running?

v0.9

Ok, we've been digging deep on this, and determined it's most likely the server node checkins causing the memory leak. It appears the go routines are not being garbage collected. We're still digging in further to get it fixed.

nice... sounds lkke my double network node issue may be a symptom of that exactly

We still have yet to encounter that. It happens even when restarting the server? Are you running with host networking? And where is the server running?

Hi! I'm having trouble accessing the dashboard at

http://dashboard.nm.X-X-X-X.nip.io/

(with my actual IP filled in for the X's) after a fresh install using the quick install script on an Ubuntu 20.04 digital ocean droplet. I don't think it's a firewall issue:

status: active

To                         Action      From
--                         ------      ----
OpenSSH                    ALLOW       Anywhere                  
443                        ALLOW       Anywhere                  
53                         ALLOW       Anywhere                  
51821:51830/udp            ALLOW       Anywhere                  
443/tcp                    ALLOW       Anywhere                  
53/udp                     ALLOW       Anywhere                  
53/tcp                     ALLOW       Anywhere                  
OpenSSH (v6)               ALLOW       Anywhere (v6)             
443 (v6)                   ALLOW       Anywhere (v6)             
53 (v6)                    ALLOW       Anywhere (v6)             
51821:51830/udp (v6)       ALLOW       Anywhere (v6)             
443/tcp (v6)               ALLOW       Anywhere (v6)             
53/udp (v6)                ALLOW       Anywhere (v6)             
53/tcp (v6)                ALLOW       Anywhere (v6)

any ideas what the issue is?

it should be https not http, but assuming that isn't working, what error do you get?

welp, I'm an idiot, thank you very much

np, I prefer the easy problems 🙂

I'm running with the default docker setup, aside from the memory limit that I set on the netmaker container (which didn't work ~ I think the container hit the memory limit and then rebooted without reconnecting to the database). Caddy is the only thing with host networking running

server is running on a Debian ec2 instance in AWS

if you delete the erroring netmaker node, does it recreate automatically?

Hi, new setup day for me and I've hit this error

ERROR: for netmaker  Cannot start service netmaker: error while mounting volume '/var/lib/docker/volumes/netmaker_sqldata/_data': failed to mount local volume: mount ~/netmaker/netmaker_sqldata:/var/lib/docker/volumes/netmaker_sqldata/_data, flags: 0x1000: no such file or directory

In docker-compose.yml I have the binds set up like this-

volumes:
  sqldata:
    driver: local
    driver_opts:
      type: none
      device: ~/netmaker/netmaker_sqldata
      o: bind
  dnsconfig:
    driver: local
    driver_opts:
      type: none
      device: ~/netmaker/netmaker_dnsconfig
      o: bind

See anything wrong? Should I be using a full path?

fixed by changing to full path, also needed to rm a previously created volume that now didn't match...

hi everyone, i used the quick install guide to install netmaker in a amazon ec2 vm as the primary. I have two raspberry pi at home, I can add both as node but they can't ping each other. I can ping the primary node from each pi but each pi can't ping to each other. Maybe I'm missing something in the config to let local hosts talk to each other?

Normally netmaker will use the local network for nodes on the same segment automatically. Not sure why it is not working for you

If you run

wg show

on one of the pi, what does it show as the endpoint of the other pi?

it shows that the endpoint of the other pi is the ip of that pi

The local lan?

right, local lan ip

not the wg node ip

I don't understand why ping would work then.