Active IQ Unified Manager Discussions

netapp harvest config metrics collection

SrikanthReddy
2,950 Views

 

 

for some of protocols seeing only write metrics and no read ,

Assuming these will be created when  read i/o is greater than 10 minimum limit defined in config ?

 

for below where do we  enable these options 

[2015-10-20 13:37:09] [NORMAL ] [smb2:node] Collection of object not enabled; skipping

[2015-10-20 13:37:09] [NORMAL ] [smb2:vserver] Collection of object not enabled; skipping

 

 

netapp harvest

netapp-harvest

p.s getting error when using netapp-harvest or harvest as labels

1 ACCEPTED SOLUTION

madden
2,915 Views

Hi,

 

You should see read/write IOPs regardless of IO level.  To have latency for read/write these require at least the threshold of IOPs (default is 10) to make sure they have enough of a base to be accureate.  My guess is either (a) you don't any read IOPs (b) it is related the perspective you are looking at or (c) your disk on the graphite server is full and the read db files can't be created.  For (b) you will see some of the info labeled as "frontend" and some as "backend" in the row headers in Grafana.  The "frontend" is where the protocol specific IOP is translated into a WAFL IOP and is where we have lots of details related to the protocol (protocol op types for example).  The "backend" is where we process the IOP in wafl and we have only read/write/other types.  So maybe you don't have LIFs with reads on the node you are looking at.

 

The SMB2 stuff being disabled is default because a summarized level of this info is enabled already under cifs object.  Only if you really want more details (like the dozens of SMB op types individually) would you want to enable these.  You would do that by using a custom config template and toggling them to be enabled.  This is described in the documentation.  I wouldn't recommend it though because in 99.9% of cases it's too much detail.

 

Lastly, you are correct, there is no 'label' for netapp-harvest, you have to use it as a 'tag'.

 

If you post more info on the specific dashboard and counter you are concerned with maybe I can comment more. Maybe include a screenshot too.

 

Cheers,
Chris Madden

Storage Architect, NetApp EMEA (and author of Harvest)

Blog: It all begins with data

 

 

View solution in original post

1 REPLY 1

madden
2,916 Views

Hi,

 

You should see read/write IOPs regardless of IO level.  To have latency for read/write these require at least the threshold of IOPs (default is 10) to make sure they have enough of a base to be accureate.  My guess is either (a) you don't any read IOPs (b) it is related the perspective you are looking at or (c) your disk on the graphite server is full and the read db files can't be created.  For (b) you will see some of the info labeled as "frontend" and some as "backend" in the row headers in Grafana.  The "frontend" is where the protocol specific IOP is translated into a WAFL IOP and is where we have lots of details related to the protocol (protocol op types for example).  The "backend" is where we process the IOP in wafl and we have only read/write/other types.  So maybe you don't have LIFs with reads on the node you are looking at.

 

The SMB2 stuff being disabled is default because a summarized level of this info is enabled already under cifs object.  Only if you really want more details (like the dozens of SMB op types individually) would you want to enable these.  You would do that by using a custom config template and toggling them to be enabled.  This is described in the documentation.  I wouldn't recommend it though because in 99.9% of cases it's too much detail.

 

Lastly, you are correct, there is no 'label' for netapp-harvest, you have to use it as a 'tag'.

 

If you post more info on the specific dashboard and counter you are concerned with maybe I can comment more. Maybe include a screenshot too.

 

Cheers,
Chris Madden

Storage Architect, NetApp EMEA (and author of Harvest)

Blog: It all begins with data

 

 

Public