Active IQ Unified Manager Discussions
Active IQ Unified Manager Discussions
I am seeing one cluster is only collecting fcp target stats and no fcp initiator stats
other cluster is not collecting any fcp stats
all other stats are ok ,
if both fcp initiator and target metrics should be collected
Solved! See The Solution
Hi,
I misunderstood your original question. Today for FC traffic Harvest only collects data from target ports because the drivers used for initiator and target ports (and the counters they support) are different. I had a look around and my colleague Paul Flores pointed me to the hostadapter object type for initiator stats. I checked and indeed there are bytes_read and bytes_written counters which work for FC and SAS initiator ports. I have added it to the backlog to be in the next release of Harvest. No ETA on it yet but it's on the list.
Cheers,
Chris Madden
Storage Architect, NetApp EMEA (and author of Harvest)
Blog: It all begins with data
P.S. Please select “Options” and then “Accept as Solution” if this response answered your question so that others will find it easily!
Hi,
I can think of four reasons you may have no data points:
1) the graphite db location is out of space (check df on the server)
2) there has been no traffic over that interface or node; harvest doesn't emit perf metrics with a value of 0 and the metrics db file only gets created when a metric has been received.
3) your graphite server is thottling creates of new metrics files and although the metric is being sent the server hasn't yet created those specific metrics files. As time passes the db file will get created
4) a bug somewhere. To narrow down where it is you would want to check the poller log in /opt/netapp-harvest/logs for errors, and if not helpful use the utlity in /opt/netapp-harvest/utils/ to check instances and counter values for the interfaces in question to see if the data ontap counter values are incrementing (harvest bug) or not (data ontap bug or misunderatnding of what it measures).
Hope this helps, and please select “Options” and then “Accept as Solution” if this response answered your question so that others will find it easily!
Cheers,
Chris Madden
Storage Architect, NetApp EMEA (and author of Harvest)
Blog: It all begins with data
Thanks chris ,
I did check the other options are not the issue
Our filers have
0e/0g as targets and seeing stats for these
0f/0h as initiators and not seeing stats
could you check if there should be any other fcp objects for initiators ?
used perf-counters-utility with -o (list objects ) saved in netapp_harvest_util_objects
>grep -i fcp netapp_harvest_util_objects
fcp_itn: CM object for exporting FCP Initiator-Target Nexus (ITN) performance counters
fcp_itn:lif: CM object for exporting FCP Initiator-Target Nexus (ITN) performance counters
fcp_itn:port: CM object for exporting FCP Initiator-Target Nexus (ITN) performance counters
fcp_itn:vserver: CM object for exporting FCP Initiator-Target Nexus (ITN) performance counters
fcp_lif: CM object for exporting FCP LIF performance counters
fcp_lif:node: CM object for exporting FCP LIF performance counters
fcp_lif:port: CM object for exporting FCP LIF performance counters
fcp_lif:vserver: CM object for exporting FCP LIF performance counters
fcp_port: CM object for exporting fcp_port performance counters
fcp_port:node: CM object for exporting fcp_port performance counters
for each of these fcp objects used below and only seeing targets but not initiator ports
-in -f (object)
I was able to create graphs for fcp metrics on targets see stats on grafana
request to add fcp targets/initiator graphs in next release grafana panels
Reference
Usage: perf-counters-utility -host <host> -user <user> -pass <pass> [-o|-in|-c|-d] [-f <family>] [-n <name> | -u <uuid>]
PURPOSE:
Collect performance data from Data ONTAP's cDOT performance counter
subsystem that uses a hierachy of: object-instance-counter
ARGUMENTS:
Required:
-host Hostname to connect with
-user Username to connect with
-pass Password to connect with
Required (one of):
-o Display object list
-in Display instance list
-c Display counter list
-d Output counter data
Required with -in, -c, -d:
-f <family> Object family
Required (one of) with -d:
-n <name> Name of instance to graph
-u <uuid> Name of UUID to graph
Optional:
-h Output this help text
-v Output verbose output to stdout
Hi,
I misunderstood your original question. Today for FC traffic Harvest only collects data from target ports because the drivers used for initiator and target ports (and the counters they support) are different. I had a look around and my colleague Paul Flores pointed me to the hostadapter object type for initiator stats. I checked and indeed there are bytes_read and bytes_written counters which work for FC and SAS initiator ports. I have added it to the backlog to be in the next release of Harvest. No ETA on it yet but it's on the list.
Cheers,
Chris Madden
Storage Architect, NetApp EMEA (and author of Harvest)
Blog: It all begins with data
P.S. Please select “Options” and then “Accept as Solution” if this response answered your question so that others will find it easily!