I subscribed few hundreds RIC (stock + options) for MARKET_PRICE & MARKET_BY_PRICE.
Running fine and after an hour or 2, it stopped with following message from log.
Using EMA Java 1.2.1 on windows 7.
Oct 23, 2018 10:39:18 AM com.thomsonreuters.ema.access.ChannelCallbackClient reactorChannelEventCallback WARNING: loggerMsg ClientName: ChannelCallbackClient Severity: Warning Text: Received ChannelDownReconnecting event on channel Channel_2 RsslReactor @7744ad0d RsslChannel @2bb02195 Error Id 0 Internal sysError 0 Error Location null Error text SocketChannel.read returned -1 (end-of-stream) loggerMsgEnd Oct 23, 2018 10:39:23 AM com.thomsonreuters.ema.access.ChannelCallbackClient reactorChannelEventCallback INFO: loggerMsg ClientName: ChannelCallbackClient Severity: Info Text: Received ChannelUp event on channel Channel_1 Instance Name Consumer_2_1 Component Version ads3.2.2.L1.linux.tis.rrg 64-bit loggerMsgEnd Oct 23, 2018 10:39:38 AM com.thomsonreuters.ema.access.LoginCallbackClient rdmLoginMsgCallback WARNING: loggerMsg ClientName: LoginCallbackClient Severity: Warning Text: RDMLogin stream state was changed to suspect with status message username <not set> usernameType <not set> State: Open/Suspect/None - text: "" loggerMsgEnd
The “SocketChannel.read returned -1 (end-of-stream)” indicates a failure condition, often that the connection is no longer available. Typically, the disconnection is from the server side.
Server log indicates that the disconnections were due to buffer overflow condition on server side. The Buffer Overflow condition generally occurs once Consumer application cannot keep up with the amount of data that it received, so server needs to store data in its outbound buffer and disconnect the Consumer client once the buffer exceeds limitation.
To avoid this slow consumer condition, you should first minimize the time for processing incoming data in the callback methods i.e. onRefreshMsg(..) and onUpdateMsg(..). Avoid accessing to database and/or other blocking calls which typically consume significant time. The application can also create an separated thread for data processing. For example, once the application processes incoming events in the callback methods, it decodes data and put it to an internal queue. The new thread then get data from the queue and process it.