Turbonomic

 View Only

  • 1.  Problem to add MSSQL Target to Turbonomic

    Posted Thu November 21, 2024 09:58 AM

    Hi,

    I have a question about adding a MS SQL Server target to Turbonomic. The SQL Server is an on-prem server and we try to add him, but it does not work. The Permissions on SQL Server side are set and the connection to the server in general works. We can see the login on the SQL Server side. But unfortunately the target always shows an error like: "There are zero target entities or all entities are inaccessible/idle". 

    We were able to reproduce the problem with a comparable server. However, unlike the original, this is a VM. We were able to temporarily change the configuration from Hostname/IP to Scope. After this change, the rediscover was successful and the connection worked. As soon as the configuration is changed back to the host name, the rediscover ends with an error message.
    As the actual SQL server is on-prem, we cannot use the scope variant there. 
    Is anyone familiar with this behavior and does anyone have a solution?
    best regards Karen


    ------------------------------
    Karen Schuster
    Technology Advocate IT Automation & Observability
    SVA System Vertrieb Alexander GmbH
    Wiesbaden
    ------------------------------


  • 2.  RE: Problem to add MSSQL Target to Turbonomic

    Posted Fri November 22, 2024 08:50 AM
    Edited by Julio Riquelme Fri November 22, 2024 08:55 AM

    Yes, I have noted that.

    My guess: 

    • Turbonomic is oriented to optimize virtual workloads.
    • So if your SQL server is virtual, it will optimize it as an standard virtual machine (say, it will keep suggesting you to add more memory over and over). 
    • Adding it also as an SQL target, Turbonomnic will identify it as such, so the "non-sense" actions will not longer appear. Plus, it will also check the connections and some other actions at the engine level.  
    • If you have an APM (Dynatrace or similar), it might add the SQL engine to the supply chain (not sure it will always do), but I dont think you will have many "optimize bare metal" actions. An APM also will let you know code-level issues that are the most common issues for an engine to perform poorly.

    I guess that optimization will be more easily achieved leveraging your APM options

    Maybe the wise guys here can add more insight.

    Regards



    ------------------------------
    Julio Riquelme
    ------------------------------

    Original Message


  • 3.  RE: Problem to add MSSQL Target to Turbonomic

    Posted Mon November 25, 2024 04:40 AM

    Karen, am I taking from this that the intended SQL server is running on a bare metal host i.e. not virtualised? If so, then it won't work trying to add the SQL server to Turbo as Turbo only works on virtualised infra or the apps running on it. This is for valid reasons as the analytics work on a economic model of buyers and sellers and so if the object hosting the application is not virtualised then Turbo knows nothing about it and so there is no "seller" in the market analytics. This explains why it works when you add a VM running a SQL server. Having APM will not help as I think the parts of the topology import from the APM where Turbo does not know of the underlying host will be thrown away.

    The only exception to this is with K8s environments that run on bare metal as Turbo does retrieve a limited amount of information about the Nodes from the K8s API (and creates VMs in the model to represent them)

    There is some useful high level on how Turbo works here, although it doesn't explicitly say that it doesn't support bare metal hosts (although there is no target for this in the Target Configuration and so it is implied) https://www.ibm.com/docs/en/tarm/8.14.2?topic=overview-how-turbonomic-works 

    Sorry for the bad news :-) 



    ------------------------------
    Barry Howard IT Automation Technical Specialist
    ------------------------------

    Original Message