truncating DFM data

[evilensky]

Hi,

We had at some point gone from no thin provisioning and no deduplication to almost all thin provisioning and almost all virtual deduplication.

However, because there was a massive drop in appearances of utilization, and actual growth has resumed, all of DFM reporting and trending is thrown off.

How would one trim the DFM historical data to ignore the utilization prior to our newfound storage efficiency and get trending data based on just the last few months?

Thank you,

Eugene

[smoot]

Wow, excellent question. Unfortunately, I know of no way to do this. Talk to your sales rep to submit an enhancement request. It's plausible but a bit hairy to implement.

[adaikkap]

Hi Eugene,

                I think using the below command, you can purge sample from the history tables of dfm.Which will help you delete some of the growth and trend graphs for the specified period.(In your case when before turning on netapp technologies.)

[root@lnx ~]# dfm purgehistory help

NAME
    purgehistory -- purge history samples from various DFM history tables.

SYNOPSIS
    dfm purgehistory timerange -c <column-name> -t <start-timestamp>
        -T <end-timestamp> [-s] [-N <replacement-value>] <table-name>
    dfm purgehistory valuerange -c <column-name> {-m <min-purge-value> |
        -M <max-purge-value> | -m <min-purge-value> -M <max-purge-value>}
        [-s] [-N <replacement-value>] <table-name>

DESCRIPTION
    The purgeHistory command neatens a DFM history table by cleaning spikes in
    history samples.

    table-name
        Table name containing history samples.

    -c column-name
        Column name containing the sample history value.

    -t start-timestamp
        Starting timestamp value of history samples for column <column-name>.

    -T end-timestamp
        Ending timestamp value of history samples for column <column-name>.

    -m min-purge-value
        Minimum purge value of history samples for column <column-name>.
        If this is the only option given, then all history samples above
        this value are purged.

    -M max-purge-value
        Maximum purge value of history samples for column <column-name>.
        If this is the only option given, then all history samples below
        this value are purged.

    -N replacement-value
        When provided, all purged history sample values for column
        <column-name> are substituted with <replacement-value>.
        Default value is zero.

    -s
        Show the list of history samples for column <column-name> without
        cleaning them from them history tables.

[root@lnx~]

Regards

adai

[chrisanthy2]

Hi Adai, I am also looking into that command, but I am not able to define what is the exact syntax for the start and end timestamp. Can you help to provide the sample syntax of:

dfm purgehistory timerange -c <column-name> -t <start-timestamp> -T <end-timestamp> [-s] [-N <replacement-value>] <table-name>

Many thanks!