Subscribe
Accepted Solution

automatic reports in performance advisor

[ Edited ]

Hello all,

 

We are trying to follow-up the performance on our NetApp boxes.

The first step is checking the CPU and disk IOPS.

 

With operations manager, I was able to get the CPU graph automatically in our report (using wget to download the biggraph).

But the disk performance report in operations manager is a bit basic.

 

Using the NetApp management console, we get nice images from performance advisor. But to get these in a report, someone has to group through all the disks, and saving the image this way.

Is it also possible to get images from performance manager in an automatic way?

 

Kind regards

Re: automatic reports in performance advisor

It's not possible to do automatically.  But using following CLI you can get the data for the each object.

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 ]

Ex:

dfm perf data retrieve  -o 1232 -V "Disk Summary View" (For multiple objects you have to give multiple objects in this CLI or write a script using this cli)

Using some graphing tool these data can be converted into a chart.
Regards,
Kishore

Re: automatic reports in performance advisor

Hi,

Thanks already, something like this was exactly what I was looking for.

But it seems that dfm perf data retrieve does not exist:

root@bru0268x:/> dfm perf data retrieve
Error: Unknown command: retrieve

Use "dfm perf help" or "dfm perf data help" for usage information.
root@bru0268x:/>

Looking what dfm perf data sub-commands exist, I get:

root@bru0268x:/> dfm perf data help

NAME
    data -- Manage performance data

SYNOPSIS
    dfm perf data list [-v | -h] [ -o <object-name-or-id> ]  [ -G <counter-group-name> ]

    dfm perf data describe [-v] <counter-group-name> <object-name-or-id>

    dfm perf data modify [ -f ] <counter-group-name>
                [ sampleRate=<sample_rate> ] [ collectionPeriod=<collection-period> ]


DESCRIPTION
    The perf data commands allow the user to list and modify
    the performance data that is collected by the server

root@bru0268x:/>

I did find a dfm perf view retrieve command, but this doens't realy gives me usefull info:

dfm perf view retrieve [ -a <appliance-name-or-id> | -g <group-name-or-id> ]
            [ -o <perf-object> ] [ -i <perf-instance> ]
            <view-name> <starttime> <endtime>

Kind regards

Re: automatic reports in performance advisor

Hi,

Sorry the data retrive CLI will be available in DFM 4.0.

perf view retrieve do same job. See below example

dfm perf view retrieve -a zidane "Disk summary view" "2009-08-20 00:00:00"  "2009-12-12 00:00:00" | more

Above CLI returns the data for given time period for all the disks in a Zidane Storage system.

Regards,

Kishore

Re: automatic reports in performance advisor

OK,

getting a bit further.

And how should I interprete this data?

View Name: Disk Summary View
View Sample Rate: 60 seconds.
View Sample Buffer: 604800 seconds.
Timestamp      huzdfas01:disk:*:user_reads
--------------------------------------------------------------------------------
14 Aug 09:35:35     1166.032959        
14 Aug 09:36:36     1754.311035        
14 Aug 09:37:35     2149.153076        
14 Aug 09:38:35     1498.050049        
14 Aug 09:39:35     1303.050049        
14 Aug 09:40:35     1439.199951        
14 Aug 09:41:35     1393.333008        
14 Aug 09:42:35     1186.883057        
14 Aug 09:43:35     1297.416992        
14 Aug 09:44:35     1321.750000        
14 Aug 09:45:35     1631.333008        
14 Aug 09:46:35     1262.833008        
14 Aug 09:47:35     1715.516968        
14 Aug 09:48:35     1291.300049        
14 Aug 09:49:35     1594.432983        
14 Aug 09:50:35     1568.217041        
14 Aug 09:51:35     1272.532959        
14 Aug 09:52:35     1365.900024        
14 Aug 09:53:35     1341.633057        
14 Aug 09:54:35     1252.282959        
14 Aug 09:55:36     1637.098022        
14 Aug 09:56:35     1552.135986        
14 Aug 09:57:35     1454.967041        
14 Aug 09:58:35     1296.782959        
14 Aug 09:59:35     1164.782959        
14 Aug 10:00:35     2459.649902        
14 Aug 10:01:35     1938.916992        
14 Aug 10:02:35     2260.366943        
14 Aug 10:03:36     1570.098022        
14 Aug 10:04:36     1631.166992        
14 Aug 10:05:35     1598.948975        
14 Aug 10:06:35     1573.883057        
14 Aug 10:07:35     1688.833008        
14 Aug 10:08:35     1644.000000        
14 Aug 10:09:35     1578.432983        
14 Aug 10:10:35     1458.817017        
14 Aug 10:11:35     1635.333008        
14 Aug 10:12:35     1545.050049        
14 Aug 10:13:35     1642.766968        
14 Aug 10:14:36     1165.508057        
14 Aug 10:15:35     1866.458008        
14 Aug 10:16:35     1314.333008       

How can I know from which disk this is?

Re: automatic reports in performance advisor

Here is the example to get the details. Here Zidane is a Storage System.

[root@wills ~]#  dfm perf view retrieve -a zidane "volume summary view" "2009-08-20 00:00:00"  "2009-12-12 00:00:00" | more

View Name: Volume Summary View

View Sample Rate: 60 seconds.

View Sample Buffer: 604800 seconds.

Timestamp      zidane:lun:kingfisher_backup/datalun/datalun:avg_latency

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

20 Aug 00:01:34     0.000000

20 Aug 00:02:34     0.000000

20 Aug 00:03:34     0.000000

20 Aug 00:04:34     0.000000

20 Aug 00:05:34     0.000000

20 Aug 00:06:34     0.000000

20 Aug 00:07:34     0.000000

20 Aug 00:08:34     0.000000

20 Aug 00:09:34     0.000000

20 Aug 00:10:34     0.000000

20 Aug 00:11:34     0.000000

20 Aug 00:12:34     0.000000

20 Aug 00:13:34     0.000000

20 Aug 00:14:34     0.000000

20 Aug 00:15:34     0.000000

20 Aug 00:16:34     0.000000

20 Aug 00:17:34     0.000000

Re: automatic reports in performance advisor

Thanks,

This indeed provided me the correct info's.

Also with the "dfm perf view list" I found all different kinds of views I could retrieve.

I also found a nice tool which can make graphs from data on a linux cli.

If anyone else need such a thing : gnuplot

Kind regards

Re: automatic reports in performance advisor

hallo tthanes24,

you mentioned, you could wget the biggraph, for example cpu1m graph. Could please post me the path to get this as picture format. I only get logo.gif and others but not the graph.

Thanks Thomas