Subscribe

harvest -graphite issue for qtree

Hello,

 

I'am having problem to get qtree perf stats in graphite.

 

I have two harvest/graphite/grafana servers installed it is working for one not for the other.

 

I can see a difference in harvest logs between the two servers.

 

On the one where it is not working :

 

[2017-08-22 15:45:03] [DEBUG ] [qtree][cdot01-02:kernel:MY_Volume/My_qtree][MY_Volume/My_qtree] [cifs_ops] = [487]
[2017-08-22 15:45:03] [DEBUG ] [qtree][cdot01-02:kernel:MY_Volume/My_qtree][MY_Volume/My_qtree] [instance_name] = [MY_Volume/My_qtree]
[2017-08-22 15:45:03] [DEBUG ] [qtree][cdot01-02:kernel:MY_Volume/My_qtree][MY_Volume/My_qtree] [instance_uuid] = [cdot01-02:kernel:MY_Volume/My_qtree]
[2017-08-22 15:45:03] [DEBUG ] [qtree][cdot01-02:kernel:MY_Volume/My_qtree][MY_Volume/My_qtree] [internal_ops] = [0]
[2017-08-22 15:45:03] [DEBUG ] [qtree][cdot01-02:kernel:MY_Volume/My_qtree][MY_Volume/My_qtree] [nfs_ops] = [0]
[2017-08-22 15:45:03] [DEBUG ] [qtree][cdot01-02:kernel:MY_Volume/My_qtree][MY_Volume/My_qtree] [vserver_name] = [svm03]

 

On the one it is working :

 

 

[2017-08-22 17:15:02] [DEBUG ] M= netapp.perf.DE0.cdot02.svm.svm09.vol.My_Vol.qtree.My_Qtree.cifs_ops 0 1503414900
[2017-08-22 17:15:02] [DEBUG ] M= netapp.perf.DE0.cdot02.svm.svm09.vol.My_Vol.qtree.My_Qtree.nfs_ops 0.337792642140468 1503414900
[2017-08-22 17:15:02] [DEBUG ] M= netapp.perf.DE0.cdot02.svm.svm09.vol.My_Vol.qtree.My_Qtree.internal_ops 0 1503414900

 

 

Here is the an extract of netapp-harvest.conf 

 

[uk0_cdot02#qtree]
hostname = cdot01
site = DE0
graphite_root = netapp.perf.DE0.cdot01
graphite_meta_metrics_root = netapp.poller.perf.DE0.cdot01#qtree
template = my_compagny-qtree.conf
instance_update_freq = 86400
data_update_freq = 300
graphite_port = 2003

 

 

here is also the template my_compagny-qtree.conf

 

##
## cDOT sample configuration template compatible with cDOT 8.3.2
##
%poller = (
## cDOT sample configuration template for qtree (requires 8.3.2 or newer)
##
'qtree' =>
{
counter_list => [ qw(instance_name instance_uuid vserver_name
cifs_ops nfs_ops internal_ops
) ],
plugin => 'cdot-qtree',
graphite_leaf => 'svm.{vserver_name}.vol.{instance_name}',
enabled => '1'
}

);

##
## Counter definition overrides
##
## Some counter definitions in Data ONTAP are defined incorrectly. This section allows overriding the
## metadata config from Data ONTAP enabling standardized processing and calculation in harvest.
##
%override = (
'cifs:node' =>
{
cifs_op_count => { properties => rate, unit => per_sec }
}
);

 

 

i compared the configurations files between the 2 servers, the controllers are 8.3.2P6.

 

Any help is welcome !Smiley Happy

Thanks

Florent

Re: harvest -graphite issue for qtree

Hi @rodillon

 

Harvest does not submit metrics that have a value of '0'.  It also does not submit latency metrics if the related IOP counter is less than 10 per second as at very low IO rates the latency metrics are not always accurate.  This 2nd situation can be controlled by adjusting the "latency_io_reqd" parameter of the poller.  More discussion of this parameter is found in the user guide.

 

Maybe check your findings with these two points?

 

Cheers,

Chris Madden

Solution Architect - 3rd Platform - Systems Engineering NetApp EMEA (and author of Harvest)

Blog: It all begins with data

 

If this post resolved your issue, please help others by selecting ACCEPT AS SOLUTION or adding a KUDO or both!

Re: harvest -graphite issue for qtree

Hi Chris,

 

You point me to an interresting option, but it was an obvious error in the pluggin file .....

That's weird i found nothing about it in harvest logs.

Any way 

 

Thanks for helping me!

 

Florent