Can you provide documentation of the steps required to configure EMA to use the b-side server for...

lancegabriel.villacrusis01

... redundancy? Or is this configured at the network layer?

Can you provide documentation of the steps required to configure EMA to use the b-side server for redundancy? Or is this configured at the network layer?

Also It looks like the login attempt was 2 hours after the disconnect. This relates to the question below

We use the Java implementation of EMA. Our understanding is that EMA will automatically reconnect to servers upon a network issue. Given the stream was reported as Open this reconnect should have occurred. It appears eventually the reconnect did occur when we started receiving prices again at 14:19 UTC. Does that mean that for this period of time the network connection was unavailable and EMA could not reconnect? Is there anything we can configure on our side such that EMA will log these reconnect attempts so we can have visibility?

Gurpreet

Hello @lancegabriel.villacrusis01,

You will have to check the RTMDS logs to see what kind of interaction was happening between the SDK and server to conclude why the failover took so long. You can also enable logging within EMA SDK to capture the raw packets when trying to replicate this issue.

Regarding EMA redundancy, you can read the following articles to apprise of available configuration options:

Resiliency in real-time feed applications

How to implement service resiliency on EMA Consumer application

Introduction to Refinitiv Real-Time SDK Warm Standby Feature

lancegabriel.villacrusis01

Thank you!

Please see the response of our RTMDS team From the server logs, I can see the following mount disconnections and reconnections that coincide with the reported time.

-----------------------------------------------

<cc1p-ecr3ads50a.1.ads: Info: Mon May 27 12:19:59.416632 2024>

RSSL disconnect from "BTEC-ADAPTIVE-PROD" at position "127.0.0.1/importer" on host "importer" and port number "48258" using application "256" of version "etaj3.7.3.L1.all.rrg|emaj3.7.3.L1.all.rrg" on channel 60.

Reason: Client application did not ping.

<END>

<cc1p-ecr3ads50a.1.ads: Info: Mon May 27 14:19:22.862210 2024>

RSSL login request from "BTEC-ADAPTIVE-PROD" at position "127.0.0.1/importer" on host "importer" and port number "49682" using application "256" of version "etaj3.7.3.L1.all.rrg|emaj3.7.3.L1.all.rrg" on channel 60.

<END>

<cc1p-ecr3ads50a.1.ads: Info: Mon May 27 14:19:22.878695 2024>

RSSL login accepted for "BTEC-ADAPTIVE-PROD" at position "127.0.0.1/importer" on host "importer" and port number "49682" using application "256" of version "etaj3.7.3.L1.all.rrg|emaj3.7.3.L1.all.rrg" on channel 60.

<END>

---------------------------------------

It seems the issue is likely isolated to the host "importer", as there are other successful mounts during the reported time.

May I know what we could advise by the question documentation of the steps required to configure EMA to use the b-side server for redundancy? Or is this configured at the network layer?

Gurpreet

https://community.developers.refinitiv.com/discussion/comment/117874#Comment_117874

The linked articles show the steps needed for implementing redundancy. Typically, it is implemented within SDK configuration, but in a rare case this can also be done with network components like a load balancer.

lancegabriel.villacrusis01

Thanks Gurpeet!