we are using RMDS instance and upgrading to EMA-API ... we are facing itemOverflowEvent issue

Manoj Bansal
Manoj Bansal Newcomer
edited January 27 in RDMS

Currently we are using RMDS servers and old rfa api. Now we are upgrading to EMA-API using same RMDS servers. which gives ItemOverflowEvent issue for snapshot request.
Same request are working fine with old API.

Can you please guide us what we are missing ?

we have provided host and port information in EMA-config file… other fields are default.

Answers

  • @Jirapongse Please help to find answer to this question

  • Hello @Manoj Bansal @dimple.shah

    Could you please give us more detail about the issue?

    • The EMA version that you are using
    • Which EMA edition that encounters the issue (C++, Java, or C#?)
    • The OS
    • The exception messages and stack trace (not just a single "ItemOverflowEvent" message)
    • How many RICs that you have sent?
    • The snippet code that creates a request message, batch or view, and calls the API registerClient() method
    • Can you replicate the issue with the EMA snapshot example? (Java, C++, or C#)
  • Hi @wasin.w

    We are currently using

    EMA-API version : 3.7.1.0
    We do not have any exception in our process logs, we are monitoring adsmon service.
    we are using Intel Xeon CPU E5-2690 v3 @2.6 GHz
    Currently we are requesting 20-25 K max in one min.

    Currently its almost the same code in java.

  • here is sample code for registerClient

    ReqMsg reqMsg = EmaFactory.createReqMsg();
            reqMsg.name("RIC");
            reqMsg.serviceName("IDN_SELECTFEED");
            reqMsg.interestAfterRefresh(false);
            reqMsg.domainType(DomainTypes.MARKET_PRICE);
            reqMsg.priority(1, 1);

            long handle = consumer.registerClient(reqMsg, this, <Closure object>);

  • Hello @Manoj Bansal

    Thank you for the information.

    The given code looks valid. It seems you have sent a single RIC snapshot request to the ADS server multiple times, Am I correct?

    About the "We do not have any exception in our process logs, we are monitoring adsmon service." statement, do you mean the ItemOverflowEvent error showed on the ADS side, not the API or application log?

    If so, I strongly suggest you contact the ADS support team to verify the issue. You can contact the team via https://myaccount.lseg.com/en/productsupport website.

    contact_ads_support.png

    I hope this information helps.

  • Hi @wasin.w

    "The given code looks valid. It seems you have sent a single RIC snapshot request to the ADS server multiple times, Am I correct?" —- Yes

    do you mean the ItemOverflowEvent  error showed on the ADS side, not the API or application log —— Yes

    let me reachout to product support

  • Jirapongse
    Jirapongse ✭✭✭✭✭

    @Manoj Bansal

    It may relate to the request throttle which is in RFA. The following is the definition throttle in RFA.

    If an OMM client application requests a large number of items at the same time, requests could overflow the outgoing channel. This could cause the channel to disconnect. At the same time, the requested item images could begin to arrive. If the consuming application cannot process the images fast enough, the network channel will overflow, causing the channel to be disconnected. When the consumer tries to recover, it will again request all of the items, thereby repeating the problem. This situation is often called thrashing. RFA uses request throttling to avoid thrashing and to ensure that the consuming application behaves in a reasonable manner

    However, EMA doesn't have the request throttle. Therefore, if the application requires this feature, developers need to implement it in the applicaiton.