none
MAP tool kit on SQL Enviornment affect the performance? RRS feed

  • Question

  • Hi Folks.

    I am planning to run MAP tool kit in SQL Eniviornment so like to know few things .

    Is MAP tool kit degrade the server or application performance during that time?

    what are the other things should keep in mind before ran MAP toolkit.

     

    Wednesday, November 9, 2011 9:00 PM

Answers

  • There are 2 different bandwidth scenarios to consider, inventory and performance data collection. MAP creates a worker thread that requests information from a target machine. The number of simultaneous worker threads created is dependent on CPU and Memory availability, but MAP has a cap that prevents it from maxing out a CPU and will throttle itself if it exceeds that cap. The bandwidth used is highly variable since it is dependent on how much data is in the reply from the target machine. In network load tests done in our test lab, I see an average of 15-20 Mbps with peaks at 50-60 Mbps.

     

    This is from our FAQ:

     

    Q: What is the expected network impact when running the MAP Toolkit?

    A: When running inventory, MAP transfers an estimated 512KB to 1MB of data per machine inventoried over the network. The main source of the variance is due to the number of items in the MSI database per machine (how many applications installed, how many patches, and so on).

    When running performance metrics collection, the amount of data collected is about 2MB per machine monitored initially and about 18-20KB every 5 minutes per machine monitored. Note that an inventory of each machine targeted for performance metrics collection will be performed if that machine has not been previously inventoried in the database being used by MAP.


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.
    • Proposed as answer by Michael Switzer01Moderator Wednesday, November 23, 2011 12:51 AM
    • Marked as answer by -kit Wednesday, November 23, 2011 9:22 PM
    Wednesday, November 23, 2011 12:50 AM
    Moderator

All replies

  • Yes, there is a significant performance impact when running MAP. But it also depends on how many machines you are inventoring. If you pull data from say a few thousand machines the network will be you bottleneck as it will use 100% of available bandwidth from the machine you run the inventory and if you have the SQL on the same machine it will also use almost all CPU making the machine unusable during inventory.

    Note that if you use SQL Server Express 2008 R2 the database size limit is 10GB which is good only for inventoring maximum ~10000 (plus/minus a few thousand) machines.


    Blogging about Windows for IT pros at www.theexperienceblog.com
    Friday, November 18, 2011 1:39 PM
  • Thank you for you detail information Andreas.

    I will go for Enviornment wise like i will try on testing enviornment first and will see results than will try for other servers.

    Thanks Again I got the info which i was wondering.


    Ankit Shah SQL Server DBA
    Friday, November 18, 2011 2:20 PM
  • There are 2 different bandwidth scenarios to consider, inventory and performance data collection. MAP creates a worker thread that requests information from a target machine. The number of simultaneous worker threads created is dependent on CPU and Memory availability, but MAP has a cap that prevents it from maxing out a CPU and will throttle itself if it exceeds that cap. The bandwidth used is highly variable since it is dependent on how much data is in the reply from the target machine. In network load tests done in our test lab, I see an average of 15-20 Mbps with peaks at 50-60 Mbps.

     

    This is from our FAQ:

     

    Q: What is the expected network impact when running the MAP Toolkit?

    A: When running inventory, MAP transfers an estimated 512KB to 1MB of data per machine inventoried over the network. The main source of the variance is due to the number of items in the MSI database per machine (how many applications installed, how many patches, and so on).

    When running performance metrics collection, the amount of data collected is about 2MB per machine monitored initially and about 18-20KB every 5 minutes per machine monitored. Note that an inventory of each machine targeted for performance metrics collection will be performed if that machine has not been previously inventoried in the database being used by MAP.


    Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread.
    • Proposed as answer by Michael Switzer01Moderator Wednesday, November 23, 2011 12:51 AM
    • Marked as answer by -kit Wednesday, November 23, 2011 9:22 PM
    Wednesday, November 23, 2011 12:50 AM
    Moderator