Hello,
I've encountered a certain problem with the recording service while recording and replaying data.
I've set-up a DataWriter that sends samples of a certain unkeyed instance, with the recording service up. Later on, I close the DataWriter and try replaying from the recording to a DataReader. By doing match analysis in the Analyzer, I see that the sample sent from the DataWriter (the replay), arrived with key, while the DataReader expects to receive no key. I'm using a very basic replay configuration that doesn't define anything concerning keyed/unkeyed instances whatsoever. Also there is a zero chance of this instance being sent from other DataWriters.
Is there any chance the problem is in the recording/replaying service or their configuration?
Thanks
Hi,
This is a very odd situation. I do not see how recording/replay service can be doing this. Let me break this down into a set of questions/observations:
Given all this the only possible explanation I can think of is that you actually had both an unkeyed and a keyed DataWriter publishing to the same Topic in your system when you run Recorder. Somehow either because Recorder was configured via XML with a keyed type, or if not because it discovered the Keyed DataWriter first, Recorder ended up creating a Keyed DataReader to record the data and then played it back as keyed...
Gerardo
Hey Gerardo, thanks for answering so fast
To clear things up and answer your questions: I don't denife the data-types in recorder config xml, I define only the topics to record.
Moreover, I've defined a new topic just to make sure I'm the only one sending and receiving data on that topic, and removed all other topics from recorder config xml. Also, while recording the data, I ran an Analyzer and saw that the DataWriter and the recorder's DataReader had a match with no key. Which makes it even more baffling that the replayer's DataWriter behaves differently from the recorder's DataReader... Is it possible that somehow the recorder DataReader received a keyed typecode from another topic that uses the same instance even though in Analyzer I saw an unkeyed match?
Thanks,
Michael
Hi Michael,
That is indeed a weird behaviour. However, we did fix an issue when getting the keyed status of the type-code in Replay and it will be available in the next general-access release.It has also been fixed in one of the patches available for Support customers.
What type of data are you sending? Is it valuetype?
Thanks,
Juanlu
Hey Juan,
The data is valuetype. Also, I'm using version 4.5f, is there a patch available for it? (I am a Support customer, could you please point me to the patch?)
Thanks
Hi Michael,
I looked you up in Salesforce and I couldn't find your account. Could you send an email to Support (support@rti.com) and ask if they can enable the RTI Recording Service 5.0.0.4 patch for you?
Thanks,
Juanlu
I've ran into a bit of a problem installing the Recorder on linux. Running ./RecordingService-5.0.0.4-linux-installer.run doesn't seem to install it or produce any output. Is it supposed to be installed by other means?
Thanks,
Michael
I'd just like to note that this issue was resolved by installing 32-bit packages of glibc and libX11