ChannelDownReconnecting event often happens with EMA Java

Takuya Miyamura

We are often experiencing "ChannelDownReconnecting event" like follows.

*=======================================================*
2018-08-06 09:00:08.947 [pool-4-thread-1] WARN c.t.e.a.OmmConsumerImpl - loggerMsg
ClientName: ChannelCallbackClient
Severity: Warning
Text: Received ChannelDownReconnecting event on channel Channel_1
RsslReactor @571db47e
RsslChannel @3034be56
Error Id 0
Internal sysError 0
Error Location null
Error text SocketChannel.read returned -1 (end-of-stream)
loggerMsgEnd
*=======================================================*

This event causes "channel out of buffers" error when reconnecting and stops our application.


Can someone please tell me:
1. What is this after all?
2. What would cause thie event?
3. Is there any way to avoid it?

umer.nalla

Hi @Takuya Miyamura

Can provide more details on when and how often you are seeing this issue? Is it on startup of your app, or when it has been running for a while?

How many RICs are you consuming and how many instances of OmmConsumer does your application use to consume these RICs?

Also, does increasing the number of output buffers resolve the issue or just delay the occurence?

e.g. increasing to 5000 for your Channel config in the EMAConfig.xml file

<ChannelGroup>
    <ChannelList>
        <Channel>
            <Name value="Channel_1"/>    
            ...
            <GuaranteedOutputBuffers value="5000"/>
            <Host value="hostname"/>
            <Port value="14002"/>
                ...
        </Channel>

Please see the EMA Config guide for further details.