Open MarcoJassy opened 2 days ago
Hi @MarcoJassy,
the participant's configuration states taht the registry's URI is silkit://0.0.0.0:8501
. That can't be right, that is the any IP address for listening sockets. You should put the IP address of the host running the registry there.
With WSL there's some subtle port forwarding and magic networking involved. You could try starting your registry on the WSL side:
sil-kit-registry --listen-uri silkit://0.0.0.0:8501
The use of the any-address is correct here. Then you need to point your Windows participant to the localhost on port 8501:
Middleware:
RegistryUri: silkit://localhost:8501
Let me know if that helps.
Hello @MariusBgm, Thank you for your reply. I forgot to mention that on this test all participant including the registry are in the same WSL. I tried to change the registry URI of the participants to localhost:8501 but than I received the same 'no delay' error already when I connected the first participant. I tried to change the ListenUri of the registry from 0.0.0.0:8501 to localhost:8501 and the result was the same as the first test. The 'no delay' error appeared when I tried to connect the second participant.
Here is the yaml file of the registry
SchemaVersion: 1
ListenUri: silkit://localhost:8501
Logging:
Sinks:
- Type: Stdout
Level: Info # Trace | Debug | Warn | Info | Error | Critical | Off
The new yaml of the participant 1
schemaVersion: 1
Description: CANoe Configuration
Middleware:
RegistryUri: silkit://localhost:8501
ConnectAttempts: 1
TcpNoDelay: true
AcceptorUris: [tcp://0.0.0.0:8502]
RegistryAsFallbackProxy: true
Logging:
Sinks:
- Type: Stdout
Level: Debug
The participant 2
schemaVersion: 1
Description: CANoe Configuration
Middleware:
RegistryUri: silkit://localhost:8501
ConnectAttempts: 1
TcpNoDelay: true
AcceptorUris: [tcp://0.0.0.0:8504]
RegistryAsFallbackProxy: true
Logging:
Sinks:
- Type: Stdout
Level: Debug
@MarcoJassy i cannot reproduce the nodelay errors, with your configs. Could you provide some more infos:
uname -a
or lsb_release -a
in the WSL shell.wsl --list --verbose
in a CMD/powershell prompt.~
just re-read your initial comment, where you state the versionfor what it's worth, I'm on a WSL version 2 running an Ubuntu 20.04 image.
Well that is strange @MariusBgm. I am using windows 10 with WSL2 and Ubuntu 22.04. My SIL Kit version is 4.0.42.
Well that is strange @MariusBgm. I am using windows 10 with WSL2 and Ubuntu 22.04. My SIL Kit version is 4.0.42.
@MarcoJassy I just attempted with SIL Kit Version 4.0.42, and it also works. I tested using the SilKit Ethernet Demo. Did you use one of the SIL Kit Demos? If not, could you test with a bundled SIL Kit Demo with your configurations and report back? Or if you have custom code, are your participants using boost::asio themselves in their code?
@MariusBgm, I started the registry of my test and than run SilKitDemoEthernet with the configuration yaml of my test as an EthernetWriter. The problem still showed up. Am I missing something in this setup?
Hello, I don't seem to be able to connect two participants in WSL with the TcpNoDelay option set to true. I receive the messages bellow every time I connect the second participant.
I am using windows 10 with WSL2 and Ubuntu22.04.
Here is the configuration yaml of the participant:
Thanks in advance.