BlueArc/Hitachi HNAS monitoring with cacti

BlueArc NAS (now part of Hitachi Data Systems) or HNAS, provide a web gui administration, like any other appliance… but it do not let one track filesystem changes nor are they integrated with the rest of my monitoring system… unless I integrate it with cacti.

As I didn’t find no developed templates for this devices, I wrote them from scratch.
This guide on cacti’s site is quite useful: Walkthrough: A New SNMP Data Query.

It guides you to create the data query and associated xml description file, to tell cacti how to retrieve data; the data template, which incorporates the needed output parameters as “data sources” and tells cacti how to store data, and the graph template, to tell cacti how to present the data.

The three concepts become somehow more clear once one have used them to construct a new query and its associated graphs.

The complete snmp MIB description for these devices can be found here: BLUEARC-SERVER-MIB.

Fortunately, cacti can make regexp parsing of snmp data, in order to retrieve some complex indexing. In this case, the so called index of the filesystem entries is composed not by the last one number of the corresponding OID, but for the two last numbers… In the xml this can be described with:


On the graph template, a correct use of cacti variables can automatically fill the name of the filesystems on the graphs by making Title equals “|host_description| – |query_hnasStorageDescr|“, where “hnasStorageDescr” was the name used on the xml file referred to in the data query.


Filesystem space is reversed in the graphs: occupation (red) grows down, and free space (blue) grows up. This is because snmp does not tell occupied space, but free and total space: so I graph the total space, and painted over it, the free space.

All this info was necessary to construct the filesystem monitoring. Once it was done, I thought that cpu monitoring was so easy that I couldn’t help but to also construct it. In this case, indexing isn’t necessary as there’s just one value to read, so a more direct approach based on a single OID suffices: the data query isn’t necessary, and the OID ( is included directly in a custom new data template.

After all the steps are complete, I exported the templates and… voilà, a brand new archive to graph Bluearc devices.

Here you can download the cacti templates for BlueArc/Hitachi NAS: on this cacti forum thread.

And here is an example of cpu and filesystem graphs. Note that space augments downwards:

cacti_hnas_example_1 In this graph the Auto-Expansion feature of these devices is shown as two steps of added space when the free space (blue) was reduced:


Once one have all the filesystems monitored, the “Aggregate plugin” can stack them all to form beautiful graphs like these: the first one with the reserved filesystems space stacked as positive values, and the free space stacked as negatives, so in just one look all info is easily shown.

cacti_hnas_example_aggregate_2 cacti_hnas_example_aggregate_1

These are the (included in a README.txt file) instructions to use the templates:

Templates to monitor a BlueArc NAS (now part of Hitachi Data Systems).
For example: HNAS 3080, 3090, 4040, etc


* File systems monitoring:
Copy this file :
to <path_cacti>/resource/snmp_queries/

* CPU monitoring:


Activate SNMP on your BlueArc device.

Now, on cacti:
Create a “Generic SNMP-enabled Host” device, and just add the:
* “SNMP – HNAS CPU Utilization” Graph Template
* “SNMP – Get HNAS Filesystem Information” Data Query

On “Create Graphs for this Host”, select “SNMP – HNAS CPU Utilization”
and the appropriate Filesystems you want to graph.

NOTE: space is reversed: occupation (red) grows down.
This is because snmp does not tell occupied space, but free and total space.

NOTE 2: cacti “ucd/net *” templates will not graph anything on these devices.

NOTE 3: default cacti templates also allows for some network monitoring.


2 thoughts on “BlueArc/Hitachi HNAS monitoring with cacti

  1. Pingback: monitoring MariaDB MaxScale with cacti | circulos meos

  2. Pingback: monitoring php-pfm threads with cacti (via snmp) | circulos meos

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s