topic OM query dfm perf data in Active IQ Unified Manager Discussions

OM query dfm perf data

martin_pauke — Thu, 05 Jun 2025 06:45:45 GMT

Hi,

is it possible to get more than one measurepoint with one query (-C aggregate:user_writes,user_reads...) (this example didn,t work).?

i use this :

dfm perf retrieve -o file:aggrname -C aggregate:user_writes -M Month -s 3600

but i need all existig information like user_writes user_reads total_transfers.

Regards Martin

OM query dfm perf data

rbalaji — Fri, 16 Sep 2011 10:56:33 GMT

You should specify the counters in comma separated format as (-C aggregate:user_reads,aggregate:user_writes,aggregate:total_transfers).

Object name was missing before user_reads in you example.

OM query dfm perf data

bipulm — Fri, 16 Sep 2011 11:03:07 GMT

One more thing is your command is wrong. It should be like dfm perf data retrieve -o <object_name> -C <counter_name> ...

You can see the usage details by using the command like

$dfm perf data retrieve help

NAME
retrieve -- This command allows you to extract the counter data
with supported statistical calculations on them.

SYNOPSIS
    dfm perf data retrieve { [ -o object-name-or-id ... ] [ -C perf-counter ... ]
                        [ -V view-name ] }
                [ -d duration ] [ -b start-time ] [ -e end-time ]
                [ -M month ... ] [ -D weekday ... ] [ -T time-range ... ]
                [ -m statistical-method [ -P percentile-value ]
                        [ -S data-advance-method ] ]
                [ -s sample-rate ] [ -x output-format ] [ -R ]

Here it clearly specifies the object name need to be there in the dfm perf data retrieve command.

OM query dfm perf data

martin_pauke — Fri, 16 Sep 2011 11:24:10 GMT

This is the output of my problem:

This worked

smc-vm-025:~ # dfm perf data retrieve -o smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k -C aggregate:user_reads -M Sep -s 3600

Timestamp smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k:user_reads

-------------------------------------------------------------------------------

2011-09-09 14:04:43 277.800

2011-09-09 15:04:43 175.867

This not:

smc-vm-025:~ # dfm perf data retrieve -o smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k -C aggregate:user_reads,aggregate:user_writes -M Sep -s 3600

Error: Expected object:counter; found aggregate:user_reads,aggregate:user_writes.

And this Tables exists:

smc-vm-025:~ # dfm perf data retrieve -o smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k -V "Aggregate Summary View" -M Aug -s 3600

Timestamp smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k:user_reads

-------------------------------------------------------------------------------

No data available

Timestamp smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k:user_writes

-------------------------------------------------------------------------------

No data available

Timestamp smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k:cp_reads

-------------------------------------------------------------------------------

No data available

Timestamp smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k:total_transfers

-------------------------------------------------------------------------------

No data available

So, the problem still exists.

Regards Martin

OM query dfm perf data

rbalaji — Fri, 16 Sep 2011 11:33:05 GMT

The 'No data available' error could be because you are asking for Aug month data and the perf data for 'Aggregate Summary View' by default stores 1 week data.

Can you please verify against Sep month ? Also, post the sampling rate and sampling duration for the 'aggregate' object of this host using 'dfm perf data describe' CLI

OM query dfm perf data

martin_pauke — Fri, 16 Sep 2011 11:43:28 GMT

This should only a example for the existing tables.

This is the main problem:

smc-vm-025:~ # dfm perf data retrieve -o smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k -C aggregate:user_reads,aggregate:user_writes -M Sep -s 3600

Error: Expected object:counter; found aggregate:user_reads,aggregate:user_writes.

All other things are only to explain.

Regards Martin

OM query dfm perf data

bipulm — Fri, 16 Sep 2011 11:48:11 GMT

Hello,

when you specify the comma in between the counter name on the cli, still -C need to be there. Otherwise it will give you error

Correct syntax would be either

$ dfm perf data retrieve -o <object name> -C aggregate:user_reads, -C aggregate:user_writes -M Sep -s 3600

without comma separated cli,

$ dfm perf data retrieve -o <object -name> -C aggregate:user_reads -C aggregate:user_writes -M Sep -s 3600

Now try again.

Thanks,

Bipul

Re: OM query dfm perf data

martin_pauke — Fri, 16 Sep 2011 12:07:21 GMT

Ok. It works now - but --- now the next question . Is it possible tho put the output in two collums. Actual the output was generated sequentialy.

smc-vm-025:~ # dfm perf data retrieve -o smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k -C aggregate:user_reads -C aggregate:user_writes -M Sep -s 3600

Timestamp smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k:user_reads

-------------------------------------------------------------------------------

2011-09-09 14:31:44 507.900

2011-09-09 15:31:44 220.283

Timestamp smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k:user_writes

-------------------------------------------------------------------------------

2011-09-09 14:31:44 507.900

2011-09-09 15:31:44 220.283

Regards Martin

OM query dfm perf data

martin_pauke — Fri, 16 Sep 2011 12:19:38 GMT

OK. I found the solution.

dfm perf data retrieve -o smc-na001-pb.smc.xxx.net:f1_aggr0_mir_15k -C aggregate:user_reads -C aggregate:user_writes -M Sep -s 3600 -x TimeIndexed

is what i need.

Thanks for your great support

Martin

Re: OM query dfm perf data

bipulm — Fri, 16 Sep 2011 12:23:45 GMT

Hello,

"-x TimeIndexed" will give you the column by column output, but it ignores the second information on the time column of the CLI output.

Thanks,

Bipul

OM query dfm perf data

martin_pauke — Mon, 19 Sep 2011 10:25:46 GMT

Hello again,

i have another "little" problem.

I could list this information with :

dfm perf counter list filer:logical_replication_destination:*:*:*

i got for example thist output:

logical_replication_destination L5S:/vol/f5_vm_esx81/lun81->LXSAN3:/vol/f3_vm_esx81_mir/lun81 num_restarts none basic

Now i would query this Objetc L5S:/vol/f5_vm_esx81/lun81->LXSAN3:/vol/.....

now my problem:

dfm perf data retrieve -o filer:L5S/vol/f5_vm_esx81/lun81->LXSAN3:/vol/f3_vm_esx81/lun81 -C logical_replication_destination:network_read_rate -s 900 -M SEP

will not work because : is a delimiter and the Object has : in his name. What is the right command Syntax with this kind of Object?

Regards Martin

OM query dfm perf data

pagrawal — Mon, 19 Sep 2011 11:21:23 GMT

Hi Martin,

Its recommended to use object Id instead of Object name.

Please try using object Id for -o option, or you can try putting double quotes on the object name .

Thanks.

Re: OM query dfm perf data

martin_pauke — Mon, 19 Sep 2011 11:34:01 GMT

Ok, where could i get the object Id for this object? I have searched in many Documentatitions but i found no example. I know only this command to resolve the objetcs:

dfm perf counter list filer:*:*:*:*

Regards

Martin

Re: OM query dfm perf data

bipulm — Tue, 20 Sep 2011 04:40:36 GMT

Hello Martin,

You can get the object id for various instances under different object type by executing the following command

$ dfm <object-type> list -q <filer>

Like for example where daood is the filer

$dfm aggr list -q dawood

$ dfm qtree list -q dawood

$dfm volume list -q dawood

The above command output having one column is "ID" which specifies the object id corresponding to the instance.

Now this object-type should be listed when you use the "dfm help" command like if you execute the below command

$dfm help

This will list aggr, lun, volume, qtree on the output. You can try also "$dfm <object-type> list help" for details.

Thanks,

Bipul

Re: OM query dfm perf data

rbalaji — Tue, 20 Sep 2011 05:18:58 GMT

Hi,

I don't think 'dfm perf data retrieve' CLI supports unmanaged objects. 'logical_replication_destination' is an unmanaged object. Further PA doesn't collect data for this object by default.

The workaround for you could be

* Create a custom view

* Add the selected counter/instance pair (i.e., logical_replication_desination:<counter-name>:<instance-name>)

From this moment, PA shall start collecting data for the counter/instance pair.