Performance Data Retention Times

[arthursc0]

Hello,

We wish to change GLOBALLY the data retntion times for performance data.

We need to ensure we have atleast 90 days worth of performance data. How do i change this globally on OM.

Have looked at cli "dfm perf data modify" but this is just too long winded to run for each counter. I want to change default to 90 days.

Regards

Colin

[adaikkap · Accepted solution]

Use the NMC->Setup->Host->Data Collection->Edit.

Set the retention on one host and, goto. NMC->Setup->Host->Data Collection->Copy.

Using the copy apply it in one shot to all host by selecting all host.

Regards

adai

[arthursc0]

Great answer. Simple and quick. Thank you.

[adaikkap]

Hi  Arthur,

     If you would like to see a global options, pls add your case to Enhancement Request 621766

Regards

adai

[reide]

arthursc0,

Sorry, but you've got to define this 90-day retention period per counter group for at least one storage array.  Once you do that, you can then copy that array's performance collection configuration and apply it to other arrays. Assuming you're using the default counter groups, here are the commands.  Copy this into a text editor and replace <hostname> with the name of one of your storage controllers.  Then, edit <host-1>, <host-2>, <host-3>.... to be the name of the other storage controllers in your environment.   Very fast and easy.

dfm perf data modify -G system -o <hostname> -r 90d
dfm perf data modify -G disk -o <hostname> -r 90d
dfm perf data modify -G aggregate -o <hostname> -r 90d
dfm perf data modify -G ifnet -o <hostname> -r 90d
dfm perf data modify -G nfsv3 -o <hostname> -r 90d
dfm perf data modify -G nfsv4 -o <hostname> -r 90d
dfm perf data modify -G prisched -o <hostname> -r 90d
dfm perf data modify -G target -o <hostname> -r 90d
dfm perf data modify -G lun -o <hostname> -r 90d
dfm perf data modify -G volume -o <hostname> -r 90d
dfm perf data modify -G cifs -o <hostname> -r 90d
dfm perf data modify -G fcp -o <hostname> -r 90d
dfm perf data modify -G iscsi -o <hostname> -r 90d
dfm perf data modify -G vfiler -o <hostname> -r 90d
dfm perf data modify -G processor -o <hostname> -r 90d
dfm perf data modify -G perf -o <hostname> -r 90d
dfm perf data modify -G priorityqueue -o <hostname> -r 90d
dfm perf data modify -G wafl -o <hostname> -r 90d
dfm perf data modify -G qtree -o <hostname> -r 90d

dfm perf data copy <hostname> <host-1>
dfm perf data copy <hostname> <host-2>
dfm perf data copy <hostname> <host-3>
dfm perf data copy <hostname> <host-4>
and so on...

[reide]

I forgot to mention.  If you're specifing a retention period that is less than the current retention period, you'll get a warning and the command won't run. This is a safety mechanism to prevent you from accidentally deleteing your historical performance data.  In this scenario, you've got to include the -f flag to force this change as it will delete anything you've collected that is older than 90 days. If you don't care about losing data older than 90 days, just include the -f flag on each command.

[rbalaji]

Therez no straight forward command to satisfy your requirement.

I hope the below perl script will solve the purpose.

You might need to replace the values in the variables @counter_groups and @hosts in the below script as per your setup.

I have tested this in our lab on a DFM 4.0 server and it worked fine.

========================================================================================

#!/usr/bin/perl
my @counter_groups = ('system', 'iscsi', 'fcp', 'lun', 'volume', 'target');

my @hosts = ('abc123.lab.netapp.com', 'xyz123.lab.netapp.com');

my $host;
my $cg;

foreach $host (@hosts) {
    foreach $cg (@counter_groups) {
        print "Modifying the retention period for host = $host and cg = $cg\n";
        system("dfm perf data modify -f -G $cg -o $host -r 90d")
    }

}

=========================================================================================

[FLISCHEIDBRACHT]

Thx @all