gns3-gui
gns3-gui copied to clipboard
Double error messages with remote GNS3VM
Describe the bug When I configure a remote server and use that as a GNS3VM each error is shown twice.
This does not happen with devices in the local server and in normal GNS3VMs.
GNS3 version and operating system (please complete the following information):
- OS: macOS High Sierra 10.13.6
- GNS3 version 2.2.10
- GNS3VM configured as a remote server
Normal local main server:

Remote server, for demonstration I'm using a standard GNS3VM in VMware fusion as the remote server:

Remote VM as a GNS3VM:

To Reproduce Steps to reproduce the behavior:
- Configure remote server and remote GNS3VM as shown above
- Stop GNS3 and start again to have a stable environment
- Add a dynamips template in the GNS3VM
- Create a project and add the dynamips node running in the GNS3VM
- Start it
- Create an error, for example
pkill dynamipsin the GNS3VM - The error message is shown twice
Screenshots or videos

Additional context
I assume that happens, because for remote GNS3VMs two notification streams are created, one for the remote server and one for the GNS3VM.
/Users/behlers# grep notifications ~/.config/GNS3/2.2/gns3_server.log
2020-06-27 18:15:56 INFO compute.py:430 Connected to compute WebSocket 'http://172.20.1.1:3080/v2/compute/notifications/ws'
2020-06-27 18:15:56 INFO compute.py:430 Connected to compute WebSocket 'http://172.20.1.131:80/v2/compute/notifications/ws'
2020-06-27 18:16:06 INFO compute.py:430 Connected to compute WebSocket 'http://172.20.1.131:80/v2/compute/notifications/ws'
Should be fixed now.
Should be fixed now.
Sad enough, it's not fixed.
I just downloaded the nighly build 2020-07-08 of v2.2 from SourceForge and re-tested. While #3010 is fixed (thanks for that), this issue is still present.
For me that's not surprising as this issue has nothing to do with the debug function, I haven't activated debuggung during this issue.
This does not happen with devices in the local server and in normal GNS3VMs.
This issue doesn't happen always, I noticed it only when I'm using a remote GNS3 VM.
Sad enough, it's not fixed.
Must be because I haven't properly tested using a GNS3 VM. I will have a closer look then.
Please note, that normal GNS3 VMs (using a virtualization engine) don't have this issue. Only GNS3 VMs using a remote VM are affected.
I assume that happens, because for remote GNS3VMs two notification streams are created, one for the remote server and one for the GNS3VM.
Correct, I could reproduce the problem. This will need some careful refactoring since we could easily break something here.
2020-07-09 20:13:38 INFO compute.py:430 Connected to compute 'ff035757-1d5b-4451-a01d-c29c11701068' WebSocket 'http://172.16.9.128:80/v2/compute/notifications/ws'
2020-07-09 20:13:38 INFO compute.py:430 Connected to compute 'vm' WebSocket 'http://172.16.9.128:80/v2/compute/notifications/ws'
This will need some careful refactoring since we could easily break something here.
Perhaps the milestone could be shifted to 2.3 or even later. For me this is more of a nuisance than a serious problem
Perhaps the milestone could be shifted to 2.3 or even later. For me this is more of a nuisance than a serious problem
Agreed, we are in the process to work on a new back-end and this could be solved at that time.