Maximo Anywhere

Expand all | Collapse all

Maximo Anywhere separate cluster

  • 1.  Maximo Anywhere separate cluster

    Posted Fri May 15, 2020 07:35 AM
    Hi Team,

    Do we require separate JVM for Maximo anywhere 7.6.4 ? 

    Thanks & Regards,
    Santhosh

    ------------------------------
    santhosh Kumar
    ------------------------------


  • 2.  RE: Maximo Anywhere separate cluster

    Posted Fri May 15, 2020 07:40 AM
    Santosh -

    Hi.  Maximo Anywhere 764 provides the ability for you to scale - depending on the number of users, usage and applications - that are being utilized.    
    As Maximo's single-vendor mobile solution - it can easily be scaled as shown in the diagram below.  

    Thanks


    ------------------------------
    Pam Denny
    IBM Maximo Mobile Offering Manager
    ------------------------------



  • 3.  RE: Maximo Anywhere separate cluster

    Posted Mon May 18, 2020 09:24 AM
    As Pam mentions in her post, Maximo is pretty flexible and allows you to isolate specific workloads (including Anywhere) if you need it. But what is required is very subjective and dependent on the organization, to make it difficult to provide a hard and fast rule as to whether or not you should isolate functionality. As an example, if you have 1 million locations or 100 locations is a large difference on the load that it will put on the Maximo system when you retrieve lookup data. Your users may primarily work offline, so they may put a heavy load early in the morning and evening as they download the work list & synchronize back to Maximo but then during the day have a negligible impact on the Maximo system because transactions aren't being sent while they're disconnected. And if you have 500 active Anywhere users at a time, that makes it much easier to justify dedicated JVMs than if you have less than 50 users. 

    Based on what we've seen with our customer base who typically have connectivity, we've chosen to scale the UI cluster over creating dedicated JVMs for the purpose of synchronizing Anywhere data and transactions. But it is something you should evaluate based on the load you expect Anywhere to put on your system.

    ------------------------------
    Steven Shull
    Director of Development
    Projetech Inc
    Cincinnati OH
    ------------------------------



  • 4.  RE: Maximo Anywhere separate cluster

    Posted Mon May 18, 2020 03:54 PM
    Steven and Pam,
    Thanks for sharing this info.
    As we do not use MobileFirst server anymore in 7.6.4, I think the "Mobile Cluster" is the solution to avoid or minimize the impact to other UI users (compare to scale up UI cluster with more JVMs) - as the load balancer for UI users and Mobile users may hit the same JVMs that are heavily used for synchronization (during peak time). I noticed this in the prior version (7.6.2) - the heaviest load is during synchronization (so many oslc calls to MX backend) when the profile is not properly (where clause) configured to only download data within smaller area where the crew is working.

    ------------------------------
    Dianne Woodley
    ------------------------------



  • 5.  RE: Maximo Anywhere separate cluster

    Posted Wed May 20, 2020 02:02 PM
    Hello Pam and Steven - Is there any specific property in Maximo which need to be configured to create a separate Mobile cluster. For sure this cluster will not run any crons etc but are there some properties which need to be set up to enable mobile on separate cluster 

    Thanks.

    ------------------------------
    Prashant Sharma
    ------------------------------



  • 6.  RE: Maximo Anywhere separate cluster

    Posted Wed August 19, 2020 08:00 PM
    Hi Prashant,

    I wrote a doc about changeing anywhere to point to a JVM deployed with another context
    https://www.ibm.com/support/pages/node/6257621

    Could you please help to review this doc?

    Thanks

    ------------------------------
    May On
    Tech Support
    IBM
    ------------------------------



  • 7.  RE: Maximo Anywhere separate cluster

    Posted Thu October 22, 2020 12:15 AM
    Hi Steve ,

    I we separate out AnyWhere cluster from UI , do we need to have this cluster also added in service integration bus for any outbound events to trigger integrations  and also need to have SEQOUT instances running on AnyWhere

    ------------------------------
    Prashant Bavane
    ------------------------------



  • 8.  RE: Maximo Anywhere separate cluster

    Posted Thu October 22, 2020 08:16 AM
    Good question. Assuming your configuration is relatively standard for the messaging engine, you should be fine. What I mean by standard is you're using files on a file share (IE \\servername\jms\myserver) that all JVMs can access or using a database. You're not using local file paths (IE D:\JMS) as that causes fragmented JMS stores where each VM has only a portion of the messages.

    How you configure the JMS queues also matters. We recommend configuring everything at a Cell level scope (not a cluster level). When configured at a cell level, that means that every JVM in any cluster can access all the JMS queues (IE UI1 can access CQIN & SQOUT, not just SQOUT). This would mean that while the messaging engine is running on the UI cluster, your other JVMs (including Anywhere) can write messages to the SQOUT which would then get consumed. It's pretty easy to test this inside of Maximo, as testing the queues in the External System application (in the Add/Modify Queues dialog using the View Queue Data action) from the JVM you're interested (such as UI or Anywhere) will throw an error if it can't access the queue. Assuming the Anywhere JVM can access the SQOUT queue, you'll be fine without any other changes.

    ------------------------------
    Steven Shull
    Director of Development
    Projetech Inc
    Cincinnati OH
    ------------------------------



  • 9.  RE: Maximo Anywhere separate cluster

    Posted Thu October 22, 2020 09:23 AM
    Thanks Steve very much appreciated taking time and providing detailed response/s.
    Yes message engines bus and resources are set up at Cell scope and datastore is used.
    So we can pretty much aligned with what you suggested.

    Thanks again.

    ------------------------------
    PB
    ------------------------------