Node Protocol Operation Test

Users access the file system of the storage system using the access protocols, including NFS, SMB/CIFS, HTTP/HTTPS, FTP, REST API, and SWIFT. The users send I/O operation requests and receive responses from the storage nodes only through the access protocols. If any of the nodes are overloaded with too many I/O operations, the particular node will not be able to quickly serve the upcoming I/O requests. This will attribute the slowdowns in I/O request processing. When the administrators experience such slowdowns, they need to swiftly identify the following:

  • Which nodes are over-loaded with operations?
  • Which protocol is sending maximum number of requests to the node?
  • What type of operations registered the maximum latency – CIFS, NFS, or iSCSI?

This test helps administrators to find out accurate answers for the above queries. This test reports the amount of data transmitted/received while performing each operation and also reveals the time delay (if any).

Target of the test : An EMC Isilon Storage System

Agent deploying the test : A remote agent

Outputs of the test : One set of results for every operation performed on each node of the target storage system.

First-level descriptor: Node

Second-level descriptor: Protocol

Third-level descriptor: Class:Operation

Configurable parameters for the test
Parameter Description

Test Period

How often should the test be executed.

Host

The IP address of the host for which this test is to be configured.

Port

The port at which the host listens. By default, this will be NULL.

Timeout

Specify the time duration (in seconds) beyond which this test should time out in the Timeout text box. The default is 10 seconds.

Connected Protocol

By default, this text box is populated with a default set of protocols separated by semi colons. The protocols listed are: cifs;ftp;hdfs;http;iscsi;ndmp;nfs;nlm;papi;siq;smb. This implies that the eG agent monitors the protocols mentioned in this text box alone and reports metrics. You can remove one/more protocols from this text box and add them back but will not be allowed to add a new protocol.

Active Protocol

By default, this text box is populated with a semi-colon-separated list of protocols that are currently active on the target storage system. For instance, the protocols listed are: cifs;ftp;hdfs;http;iscsi;ndmp;jobd;Isass_out;nfs;nfs3;nfs4;nlm;papi;siq;smb2. This list will only include one or more protocols or the variants of the protocols mentioned in the Connected Protocol text box.

Use SUDO

By default, this parameter is set to No. This indicates that, by default, the eG agent will not require any special permissions to execute the commands. However, in some highly secure environments, this command cannot be executed directly as the eG agent install user is different from the root user who has the privileges to run all commands on the target storage system. In such cases, create a sudo user using the steps discussed in the Pre-requisites for Monitoring the EMC Isilon Storage System. Credentials of such a user should be specified in the Username and Password text boxes in the COMPONENTS page.

Set the Use Sudo parameter to Yes. This will enable the eG agent install user to execute the commands.

Measurements made by the test
Measurement Description Measurement Unit Interpretation

Total operations

Indicates the rate at which operations of this type were performed.

Operations/sec

The value of this measure reveals the workload received via each protocol to each node in the storage cluster.

Data received

Indicates the rate at which the data was received during performing this operation.

KB/sec

Compare the value of this measure across the operations to identify the operation during which the maximum amount of data was received.

Data transmitted

Indicates the rate at which the data was transmitted out during performing this operation.

KB/sec

Compare the value of this measure across the operations to identify the operation during which the maximum amount of data was transmitted.

Average time

Indicates the average time taken for performing this operation.

Milliseconds

A high value of this measure is a cause for concern, as it indicates a processing bottleneck.

Standard time deviation

Indicates the standard time deviation of this operation.

Milliseconds