I'm working with RTI connext 5.0.0 and using an older quality of service file from a previous project running connext 4.5d. It doesn't seem to have any problems running with the QoS file, but I'm getting a bizarre error on write that I haven't seen before. It seems to only happen when a reader goes down, and it leads me to believe the writer on the same topic is getting stuck trying to write to it, which didn't happen in version 4.5d.
I've attached the QoS file I'm using. Is there some new parameter I need to set to prevent this writer/reader blocking issue? If I reader dies I don't want my writer failing to write to everyone else because of it.
I checked the knowledge base, the forum, AND the API/docs, but there's zero documentation on this error or what causes it.
| Attachment | Size |
|---|---|
 qos_library.xml | 21.43 KB |
Hi MwKozak,
This error message may be caused by the small history depth you are configuring. Since you are using asynchronous publishing, samples are stored in a queue before being flushed to the network. The problem may be that before being flushed, some samples are discarded (because you only want the last 10 samples), so that error is printed.
To discard that reason, I would give a try to set a history depth bigger for the DataWriter and the DataReader.
If that doesn't fix the issue, please run the program in DEBUG mode and post here the more complete log.
Regards,
Juanjo Martin