Closed jc211 closed 1 year ago
That's weird. The output in the trace is entirely consistent with using domain id 230, and when I tried it (ddsperf -i 230 sanity
as well as with a small Python script) it all worked as expected.
What I am wondering is if in your case perhaps different Cyclone traces are overwriting each other. It is easy to get the process id in the file name:
<Tracing>
<OutputFile>cdds.log.${CYCLONEDDS_PID}</OutputFile>
</Tracing>
If you also have multiple domains in a single process,
<Tracing>
<OutputFile>cdds.log.${CYCLONEDDS_DOMAIN_ID}.${CYCLONEDDS_PID}</OutputFile>
</Tracing>
will do the trick.
That is, as long as CYCLONEDDS_PID
and CYCLONEDDS_DOMAIN_ID
are not defined as environment variables, because if the environment defines them, they take precedence.
I think this is exactly the problem. I was creating 2 domains, one for the subscriber and one for the publisher in the same program for testing purposes and I think one overwrote the other in some way. Thank you :+1:
Hi,
I'm creating a domain and a participant with id=230 on the loopback interface. I think this being done correctly because the debug data confirms the id (see below).
The problem is the participant is only visible when writing
cyclonedds ps --id 0
and nothing shows up withcyclonedds ps --id 230
I'm not sure how this is possible unless the domain id is not actually being set.