Post reply

SQL Trace file sizes

  • July 30, 2014 at 3:09 am

    #297206

    Hi there

    I am in the middle of capturing a workload to try and tune a SQL instance and was wondering what kinds of sizes people capture in terms of traces. I am only 1 day into a capture and I believe a typical workload would be a week long capture and I am already at 10GB of files.

    I am only capturing rpc_completed and sql_batch_completed.

    What sizes of workloads do other people capture and then where do you analyse them, do you have particular dedicated server for this kind of thing as at present I am looking to use my local PC. Also what rollover file sizes do people tend to use, I am currently using 1GB.

    Thanks in advance for writing about your own experiences.

  • July 31, 2014 at 12:24 am

    This was removed by the editor as SPAM

  • July 31, 2014 at 2:42 am

    #1733953

    Hi thanks for that but I was more interested in how people collect their workloads in the first place do people just filter on a particular database over a time period of what they consider a typical workload, or do people collect at the instance level and have 100s of GB of data to analyse?

  • July 31, 2014 at 10:23 am

    #1734152

    We've been using 25MB or 100MB file size for years now, for traces ranging from 10MB to 5GB in total size. Haven't had any problems with these file sizes. Years ago we experimented a bit with different sizes, and arrived at these; have not reconsidered since.

  • July 31, 2014 at 10:54 pm

    This was removed by the editor as SPAM

  • August 1, 2014 at 2:26 am

    #1734372

    Yes that was my point really, I wasn't able to get a decent workload as I collected too much information.

    I wondered if collecting workloads should be performed at the database level to keep the amount collected down, it sound like you both do? What about at table level also do people collect workloads at that level?

    I see a lot of mentioning of ensure you collect workloads, but never have I seen a detailed explanation of what that means.

  • August 1, 2014 at 8:41 am

    #1734459

    In my cases, most workloads are cross-database. So we capture activity either at a gross "everything" level, or more often, just at the specific workload level.

  • August 1, 2014 at 9:05 am

    #1734472

    I see, so you generally get a set of users to simulate some workload and capture that.

Viewing 8 posts - 1 through 7 (of 7 total)

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