I don’t hear much about this command from the NetApp user community – stats. It’s a nifty little command that allows us to monitor a lot of performance metrics at a fairly granular level. The command output can also be customized, so this might save a lot of parsing and grep-ing that goes on during our normal filer performance monitoring, charting exercises.
First looks:
san-a> stats The following commands are available; for more information type "stats help <command>" explain list start stop help show stats help - Display command help. stats list - List available objects, instances, counters or presets in system. stats explain - Explain available objects or counters in system. stats show - Show all or selected statistics with formatting. stats start - Start background statistic collection. stats stop - Stop all background operations and discard results. san-a> san-a> stats list Usage: stats list objects [-p <preset>] stats list instances [-p <preset>] [<object_name>] stats list counters [-p <preset>] [<object_name>] stats list presets - List available objects, instances, counters or presets in system. san-a>
stats uses objects, instances and counters to collect and report various performance metrics. Objects are fixed and can be considered as global counters. There are multiple instances of an object (for e.g. disk is an object and for each disk in the system, there is an instance). For each of these instances, ONTAP can collect and report a variety of performance metrics. These individual metrics are called counters.
You can see how we can get as global of a view as possible or as specific of a view as possible using the stats command.
What objects are there ?
san-a> stats list objects Objects: system disk processor ifnet nfsv3 target lun volume cifs fcp iscsi aggregate qtree quota vfiler ext_cache ext_cache_obj logical_replication_destination logical_replication_source dump ndmp hostadapter san-a>
There may be multiple instances for some of these objects. To see the instances, we can do a
san-a> stats list instances lun Instances for object name: lun /vol/luns/demolun7-C4aYk4Lggh9R /vol/luns/san-a-lin03-fcp-lun1-C4aYk4L-g3pl /vol/luns/san-a-lin03-lun1-C4aYk4Dlv0bh /vol/luns/san-a-lin03-lun2-C4aYk4Dlv0bp /vol/luns/san-a-lin03-lun0-C4aYk4Dlv0bY /vol/luns/demolun4-C4aYk4IzJb3b /vol/luns/san-a-lin03-fcp-lun0-C4aYk4F7X3ff san-a>
For each of these instances, we can get a variety of performance metrics:
san-a> stats list counters lun Counters for object name: lun read_ops write_ops other_ops read_data write_data queue_full avg_latency total_ops san-a>
Given this info, let’s go thru some “progressive metrics extraction”..
san-a> stats show volume volume:vol0:avg_latency:37.92us volume:vol0:total_ops:14/s volume:vol0:read_data:0b/s volume:vol0:read_latency:0us volume:vol0:read_ops:0/s volume:vol0:write_data:263b/s volume:vol0:write_latency:231.50us volume:vol0:write_ops:2/s volume:vol0:other_latency:5.66us volume:vol0:other_ops:12/s volume:luns:avg_latency:0us volume:luns:total_ops:0/s volume:luns:read_data:0b/s volume:luns:read_latency:0us volume:luns:read_ops:0/s volume:luns:write_data:0b/s volume:luns:write_latency:0us volume:luns:write_ops:0/s volume:luns:other_latency:0us volume:luns:other_ops:0/s volume:maska:avg_latency:0us volume:maska:total_ops:0/s volume:maska:read_data:0b/s volume:maska:read_latency:0us volume:maska:read_ops:0/s volume:maska:write_data:0b/s volume:maska:write_latency:0us volume:maska:write_ops:0/s volume:maska:other_latency:0us volume:maska:other_ops:0/s volume:maskb:avg_latency:0us volume:maskb:total_ops:0/s volume:maskb:read_data:0b/s volume:maskb:read_latency:0us volume:maskb:read_ops:0/s volume:maskb:write_data:0b/s volume:maskb:write_latency:0us volume:maskb:write_ops:0/s volume:maskb:other_latency:0us volume:maskb:other_ops:0/s volume:maskc:avg_latency:0us volume:maskc:total_ops:0/s volume:maskc:read_data:0b/s volume:maskc:read_latency:0us volume:maskc:read_ops:0/s volume:maskc:write_data:0b/s volume:maskc:write_latency:0us volume:maskc:write_ops:0/s volume:maskc:other_latency:0us volume:maskc:other_ops:0/s volume:oracledata:avg_latency:0us volume:oracledata:total_ops:0/s volume:oracledata:read_data:0b/s volume:oracledata:read_latency:0us volume:oracledata:read_ops:0/s volume:oracledata:write_data:0b/s volume:oracledata:write_latency:0us volume:oracledata:write_ops:0/s volume:oracledata:other_latency:0us volume:oracledata:other_ops:0/s volume:oraclebin:avg_latency:0us volume:oraclebin:total_ops:0/s volume:oraclebin:read_data:0b/s volume:oraclebin:read_latency:0us volume:oraclebin:read_ops:0/s volume:oraclebin:write_data:0b/s volume:oraclebin:write_latency:0us volume:oraclebin:write_ops:0/s volume:oraclebin:other_latency:0us volume:oraclebin:other_ops:0/s volume:oraclelogs:avg_latency:0us volume:oraclelogs:total_ops:0/s volume:oraclelogs:read_data:0b/s volume:oraclelogs:read_latency:0us volume:oraclelogs:read_ops:0/s volume:oraclelogs:write_data:0b/s volume:oraclelogs:write_latency:0us volume:oraclelogs:write_ops:0/s volume:oraclelogs:other_latency:0us volume:oraclelogs:other_ops:0/s volume:esx_vol1:avg_latency:0us volume:esx_vol1:total_ops:0/s volume:esx_vol1:read_data:0b/s volume:esx_vol1:read_latency:0us volume:esx_vol1:read_ops:0/s volume:esx_vol1:write_data:0b/s volume:esx_vol1:write_latency:0us volume:esx_vol1:write_ops:0/s volume:esx_vol1:other_latency:0us volume:esx_vol1:other_ops:0/s volume:aggr1_vol1_esx:avg_latency:83.33us volume:aggr1_vol1_esx:total_ops:3/s volume:aggr1_vol1_esx:read_data:0b/s volume:aggr1_vol1_esx:read_latency:0us volume:aggr1_vol1_esx:read_ops:0/s volume:aggr1_vol1_esx:write_data:20480b/s volume:aggr1_vol1_esx:write_latency:83.33us volume:aggr1_vol1_esx:write_ops:3/s volume:aggr1_vol1_esx:other_latency:0us volume:aggr1_vol1_esx:other_ops:0/s san-a> san-a> stats show volume:aggr1_vol1_esx volume:aggr1_vol1_esx:avg_latency:0us volume:aggr1_vol1_esx:total_ops:0/s volume:aggr1_vol1_esx:read_data:0b/s volume:aggr1_vol1_esx:read_latency:0us volume:aggr1_vol1_esx:read_ops:0/s volume:aggr1_vol1_esx:write_data:0b/s volume:aggr1_vol1_esx:write_latency:0us volume:aggr1_vol1_esx:write_ops:0/s volume:aggr1_vol1_esx:other_latency:0us volume:aggr1_vol1_esx:other_ops:0/s san-a> stats show volume:aggr1_vol1_esx:avg_latency volume:aggr1_vol1_esx:avg_latency:0us
OR
I can extract one particular counter from across all of my volumes
san-a> stats show volume:*:avg_latency volume:vol0:avg_latency:34.90us volume:luns:avg_latency:0us volume:maska:avg_latency:0us volume:maskb:avg_latency:0us volume:maskc:avg_latency:0us volume:oracledata:avg_latency:0us volume:oraclebin:avg_latency:0us volume:oraclelogs:avg_latency:0us volume:esx_vol1:avg_latency:0us volume:aggr1_vol1_esx:avg_latency:0us
or I can select TWO counters across all volumes
san-a> stats show volume:*:avg_latency volume:*:write_ops volume:vol0:avg_latency:38.09us volume:vol0:write_ops:2/s volume:luns:avg_latency:0us volume:luns:write_ops:0/s volume:maska:avg_latency:0us volume:maska:write_ops:0/s volume:maskb:avg_latency:0us volume:maskb:write_ops:0/s volume:maskc:avg_latency:0us volume:maskc:write_ops:0/s volume:oracledata:avg_latency:0us volume:oracledata:write_ops:0/s volume:oraclebin:avg_latency:0us volume:oraclebin:write_ops:0/s volume:oraclelogs:avg_latency:0us volume:oraclelogs:write_ops:0/s volume:esx_vol1:avg_latency:0us volume:esx_vol1:write_ops:0/s volume:aggr1_vol1_esx:avg_latency:0us volume:aggr1_vol1_esx:write_ops:0/s san-a>
This is ONLY a high level overview of the stats command. The following links will help understand the stats command fully.
Further Reading:
- A quick guide to using the stats command
stats – man page
stats – who is your Daddy and what does he do?
Make stats part of your daily scripting regimen and you’ll find this a time saver and a swiss army knife.
