Skip to main content

storage aggregate object-store profiler start

Contributors
Suggest changes
PDFs

Start the object store profiler to measure latency and throughput

Availability: This command is available to cluster administrators at the advanced privilege level.

Description

The storage aggregate object-store profiler start command writes objects to an object store and reads those objects to measure latency and throughput of an object store. This command requires two parameters - an object store configuration and node from which to send the PUT/GET/DELETE operations. This command verifies whether the object store is accessible through the intercluster LIF of the node on which it runs. The command fails if the object store is not accessible. The command will create a 10GB dataset by doing 2500 PUTs for a maximum time period of 60 seconds. Then it will issue GET operations of different sizes - 4KB, 8KB, 32KB, 256KB for a maximum time period of 180 seconds. Finally it will delete the objects it created. This command can result in additional charges to your object store account. This is a CPU intensive command. It is recommended to run this command when the system is under 50% CPU utilization.

Parameters

-node {<nodename>|local} - Node on Which the Profiler Should Run (privilege: advanced)

This parameter specifies the node from which PUT/GET/DELETE operations are sent.

-object-store-name <text> - Object Store Configuration Name (privilege: advanced)

This parameter specifies the object store configuration that describes the object store. The object store configuration has information about the object store server name, port, access credentials, and provider type.

[-object-prefix <text>] - Prefix Added to Each Object (privilege: advanced)

This parameter specifies the prefix attached to each object. There is a prefix length restriction of 80 characters. In addition to this restriction, follow any specific prefix length or character restrictions that are imposed by the cloud store they plan to run this command. Refer to the respective cloud store documentation for details.

Examples

The following example starts the object store profiler :

cluster1::>storage aggregate object-store profiler start -object-store-name my-store -node my-node