Open R290 opened 1 year ago
same issue here. OPCUA Server gets flooded with wrong session ids.
Just checked with v0.5.2 and the flooding of the server no longer occurs. @Scaredon which version are you using? The read example at https://github.com/gopcua/opcua/blob/main/examples/read/read.go has also been modified to make a distinction between different error types.
@R290 I've retested today. newest telegraf plugin with v0.5.2 goopcua. Local Build. Autoreconnect is set to true.
The server is spamed with the sequence (not sure if this is the correct order, log frequency is pretty high):
Sadly, I'm unable to reproduce this. Could you provide any information about the server and authentication method you are using?
Related to https://github.com/influxdata/telegraf/issues/11559 and possibly related to https://github.com/gopcua/opcua/issues/597 and https://github.com/gopcua/opcua/issues/434.
The errors below are generated using the slightly modified read example here https://github.com/R290/opcua/tree/autoreconnect/examples/read with the
docker-compose
in the same folder.The output with debugging disabled:
When restarting the opcua container:
It does not recover cleanly.
Normal behavior with debugging enabled:
When restarting the opcua container the debug console is flooded witht the following repeated message patten:
Via Wireshark at localhost:3000 we can also see that the server is being swamped with packages (using filter
tcp.port == 4840
). The reason Nginx is present is so we can simulate a dropped connection as well by restarting the nginx container. Behavior is similar to restarting the opcua server itself.To me it seems that flooding the server with requests on a dropped connection is definitely not as designed...