Post reply

Replays - Simulating Typical Load on a Second Instance

  • January 14, 2016 at 9:39 am

    #306923

    Good Afternoon All, I hope you're each well?

    I have a 2 Node production box that is going to be replaced with VM's (Equivalent or Higher Spec, but virtual all the same)

    I was hoping to do the following:

    Take a backup of production and restore on the newly built VM's

    Grant testers access and let them do their typical thing.

    At the same time I wanted to try and run a typical days work load against it whilst monitoring performance.

    I've never done this before using a SQL Trace, have only used it for identifying performance problems...

    Could someone point me in the right direction please?

    IF I run a workload against an instance, does it do it the same time frame as when it was captured, or will it blast through it as quick as it can?

    Do I need the distributed replay Coordinator involved in this?

    Any and all help is appreciated

    Cheers

    Alex

  • January 14, 2016 at 2:41 pm

    #1852104

    You have two options:

    1) Distributed replay

    2) RML utilities

    Both tools support replaying the workload in "synchronization mode", which is to say that queries are run at the same rate they were recorded, differently from what happens in "stress mode" (queries are run one after another, with no pauses).

    There's a quick introductory post on my blog: http://spaghettidba.com/2012/11/22/replaying-workloads-distributed-replay/

    Hope this helps

    Gianluca

    -- Gianluca Sartori

  • January 15, 2016 at 7:09 am

    #1852241

    Cheers! Most helpful... Had not even heard of RML until now.

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply