How to create a HVR Statistics report

Viewing 1 post (of 1 total)
  • Author
    Posts
  • #13944
    ggoodrich
    Keymaster

    Question:

    How to create a consolidated weekly report, including:

    • Number of records/rows replicated including a break-down by operation type (insert, update, delete)
    • Size of data sent between source and target

     

    Response:

    HVR includes both monitoring and statistics capabilities. For statistics, HVR logs detailed statistics to the various output logs which can be aggregated using our hvrstatistics utility. HVR statistics is a utility that analyzes the HVR’s output log files and returns several types of statistical information including both number of rows replicated along with the size of data captured, compression rates, and size of data processed on the replication system(s). Generating the HVR Statistics from the statistics utility can be scheduled to run from the HVR Console to produce statistics or run from the command line.

    There are basically two methods to generate statistics, which includes; 1) ad hoc graphical display against a channel or one of its job groups, or 2) statistics channel to analyze output logs to generate CSV values which are loaded into tables in your repository. The Ad Hoc approach is run on demand to display nice graphical charts based on your user-defined criteria. The scheduled approach does not generate the sexy graphics, but what it provides is data into tables which you can then use to generate your own charts and statistics for your weekly management reports. Play with both approaches to determine which option addresses your need.

    What I recommend is establish a Statistics Channel job which is scheduled once per day or once per hour or whatever interval you desire. This generates a job that will run automatically based on the interval you select. The statistics channel parses the HVR.OUT output log looking for statistical information in CSV files that are loaded into a relation table. As statistics are collected, the data is loaded in a HVRSTATS table in the repository (which needs to be created).

    Statistics Channel approach

    I have included an additional “How-to” document, named HVR Statistics Channel.pdf which provides the steps to create the HVRSTATS table, along with steps to schedule the statistics channel.

    Ad hoc approach

    Here you can see me select the schedule for the channel, right-click, and select statistics. The following window is displayed. I then selected to include/exclude what information is included in the statistical output generated. Here I selected a daily view which will aggregate all values collected since the last daily run.

    For example:

     

    You can then view the statistics as seen in the example below. You will see you can adjust which information is displayed, that may include:

    • Replicated rows
    • Transaction counts
    • Speed
    • Duration
    • Run cycles
    • Bytes
    • Compression rates
    • Latency value

    For example, below I requested statistics to be displayed includes replicated rows, transaction counts, number of cycles, and size of bytes replicated.

     

    For additional Reference materials, please review:

    • How to forum — HVR Statistics Channel.pdf – HVR Statistics in a relational format for reporting
      • This presents a how-to on collecting HVR statistics for reporting purpose
    • HVR User Guide – Reference Guides
      • See HVRStatistics, which provides a summary, syntax and example of using the hvrstatistics command line utility
Viewing 1 post (of 1 total)
  • The forum ‘How to’ is closed to new topics and replies.

© 2020 HVR

Test drive Contact us