Coreflux Cluster Resilience Test Tool

Description

The Coreflux Cluster Resilience Test Tool provides an easy way to check for payload losses, when using the cluster.

This tool creates a user input amount of clients that connect to a cluster broker that then send the payload from 0-9 in user input intervals. An extra client, that subscribes to the topics the payload is being sent to from the publishers, determines if the sequence received is valid, testing the cluster's resilience.

Prerequisites

• Cluster IP Address;

Setup instructions

1. Go to the Releases tab;
2. Download the zip file for your OS;
3. Extract the contents;
4. Run the executable file;

OR

4. Open the console on the directory of the file;
5. Run:
   On Linux

   cf_mqtt_broker_cluster_resilience_test_tool <broker_ip> <port> <topic_prefix> <nmr_of_clients> <publish_time_ms> <run_time_s> <behaviour>
   

   On windows:

   cf_mqtt_broker_cluster_resilience_test_tool.exe <broker_ip> <port> <topic_prefix> <nmr_of_clients> <publish_time_ms> <run_time_s> <behaviour>
   

Using the first method:

The user will be prompt by the console to introduce the broker's ip address, the port, the topic prefix, the number of clients, the publish time in ms, the run time in seconds and the program's behaviour:

Broker Address: 
Port:
Topic Prefix:
Number of clients:
Publish time (ms):
Run time (s): 
Select a behaviour option:

(1) Stop at the first missed payload
(2) Count the missed payload

Option:

Name	Description	Example
Address	IP address or hostname of the MQTT cluster.	127.0.0.1
Port	Port number on which the MQTT cluster is running.	1883
Topic Prefix	The prefix of the topic the clients will be sending / receiving the payload	mqtt/
Number of clients	The number of clients that are going to connect and send the payload to the cluster	10
Publish time	The time between each payload sent to the cluster, in milliseconds	100
Run time	The time the program will run for, in seconds	60
Behaviour	If the program should stop when it detects the first missed payload (option 1) or if it should run and count all the instances it happens (option 2)	2

Output

Failure

If there has been a miss on the sequence and the behaviour selected was 1, the program will stop, disconnect all the clients and prompt the user to close:

Sequence error detected: <topic>: Previous = <previous_value>, Actual = <actual_value>
Disconnecting clients...

Press any key to close

where topic is the topic the subscriber detected the miss, previous_value the last value received and actual_value the value that skipped the sequence.

If there's an exception other than a miss on the sequence, you'll get the message:

Error: <error_information>

After run time

If no sequencial payload was lost, you'll get the following message:

Finished with no payload lost. Elapsed time: <elapsed_time>

If the behaviour selected was the option 2 and there has been misses in the sequence, you'll get the following message:

Finished with a total of <number_of_misses> payload loss. Elapsed time: <elapsed_time>

where number_of_misses the count of the payloads that were skipped and elapsed_time is the time the program ran for.

To close the program you'll be prompted to press any key of the keyboard:

Press any key to close