Content Management and Capture

 View Only
  • 1.  Batch date in BatchId does not update when the next day comes.

    Posted Wed January 12, 2022 02:56 AM
    Edited by dsakai Wed January 12, 2022 03:02 AM
    Hi,
    The client is running 4 Rulerunner Servers each having 8 processes.
    So 32 processes run in parallel. Datacap is 9.1.5 ifix7.
    The client is seeing that for some reason batch date does not change when the next day comes.
    They have bee seeing "20211220.00xxxxxx" batch folders being created for more than 2 weeks.
    The serial has become something like "20211220.015000".
    I've told them to restart Datacap Server to see the batch date renewed. They did, but no change.
    Yesterday (1/11), for some reason, the batch date suddenly updated from 20211220 to 20211224.
    The batch date now stays at "20211224.00xxxxxx", not updating to 20220111.
    Does anyone know why Datacap is not updating batch date and how to fix it so each new day starts with 0 numbering.

    ------------------------------
    dsakai
    ------------------------------


  • 2.  RE: Batch date in BatchId does not update when the next day comes.

    Posted Thu January 13, 2022 06:09 AM
    Sure it could be a Datacap problem but to be sure you did everything correct
    I would check where and when the batch is created. For example if you have a VScan task which should run in background and the batch must be created in the thread you add it, then you have to check the date on this Rulerunner.
    Theoretically you can have only one VScan running in Rulerunner thread in background for an Datacap application. It could happen when the threads and the VScan were not correctly defined (see parameters for this in Rulerunner Manager) that about every minute  a new batch is created... perhaps a lot of batches on 22.11.2021 for example. Then first when all theses batches in wait will be processed only after this a new batch is created.

    Resume: check in monitor when the batches are created exactly and check the Date on DC Rulerunner and Datacap TaskManager Server.
    I hope it helps.
    dorothea



    ------------------------------
    dorothea vulcan
    dorotheavulcandorotheavulcan@yahoo.com
    ------------------------------



  • 3.  RE: Batch date in BatchId does not update when the next day comes.

    Posted Sun January 16, 2022 02:13 AM
    I checked Server Time on all Datacap Servers and Rulerunner Servers.
    They are all correct.
    I think MVScan configuration is correct as well. It doesn't cause this Batch Date issue in Dev/Test environment, which has a lot less Rulerunner concurrent processes. So I am guessing 32-48 concurrent RR run has something to do with it.
    Do you happen to know where or what messages to look in tms.log for this issue?

    ------------------------------
    dsakai
    ------------------------------



  • 4.  RE: Batch date in BatchId does not update when the next day comes.

    Posted Tue January 18, 2022 05:32 AM
    Perhaps if I run MVScan action in only one of 48 Rulerunner threads, Batch date may update normally.
    However, this configuration may most likely not be good in production.
    If that single thread dies or suspends, the whole production halts because no more input comes in.
    But since in Dev/Test environment, 4 threads each runs MVScan and batch date problem doesn't happen,
    we will consider reducing MVScan threads to a lesser number in production.
    Thank you for this suggestion.

    ------------------------------
    dsakai
    ------------------------------



  • 5.  RE: Batch date in BatchId does not update when the next day comes.

    Posted Tue January 18, 2022 11:47 PM
    According to this document below "MVScan vs VScan Ingestion Differences in IBM Datacap",
    VScan needs to be placed in only one thread. However, MVScan is not limited to one thread.
    However, 48 threads running MVScans may not be good idea.
    We are considering about reducing MVScan threads to one or two per Rulerunner Server.

    https://www.ibm.com/support/pages/mvscan-vs-vscan-ingestion-differences-ibm-datacap

    "Cause
    VScan is able to ingest the newly scanned images because it assumes it has full control over
    the ingestion folder. With no other competing threads, it waits for more images (up to SetMaxImageFiles value) to ingest.

    Answer
    MVScan assumes that there are other threads pointing to the same ingestion folder so
    MVScan will not have that same symptom. This is working as designed."

    ------------------------------
    dsakai
    ------------------------------