mosquitto

How about this test case?

Asked by Sang kyeong Nam

I test mosquitto with this environment.

server : mosquitto-0.8.3
mosquitto setting:
  retry_interval 30
  store_clean_interval 1
  max_inflight_messages 1000
  max_queued_messages 0

  autosave_interval 10
  persistence 1

receivers :
   execute 100 mosquitto_sub by bash script in one PC.

sender : using wmqtt.jar, written in Java. in one PC, but different with receivers.
  send 30 message each receiver. (total messages count is 3000)

I Test with this flow

1. Execure mosquitto.
2. Receivers connect to server with QoS 2, clean-session is false. And subscribe.
3. Disconnect all receiver. Don't unsubscribe.
4. Sender send 30 messages each receiver.
5. Reconnect all receiver when sending message is completed.

Result.
  38 Receivers fail to connect after test flow 5.
  1710 message receive successful.
  150 message retain on server.

  (38*30)+1710+150 = 3000

I want to know how to get good results.

Question information

Language:
English Edit question
Status:
Solved
For:
mosquitto Edit question
Assignee:
No assignee Edit question
Solved by:
Sang kyeong Nam
Solved:
Last query:
Last reply:
Revision history for this message
Roger Light (roger.light) said : 		#1

Hi Sang,

Could you please retry the same test with mosquitto 0.9.3? There have been quite a few changes and importantly some of the max_inflight_messages and max_queued_messages code has been fixed.

Revision history for this message
Sang kyeong Nam (nskpc1) said : 		#2

server : mosquitto-0.9.3

Result.
  38 Receivers fail to connect after test flow 5.
  1860 message receive successful.

I think what the connection failure is linux issue .
After Result, I try connect to server from the machine. But can't connect.
And I try connect to server from another machine. It's success.

I have another issue in mosquitto-0.9.3
When 100 receivers connect server, I send 30 messages to each receivers with QoS 2.
Every receiver has receive duplicate message.

Received PUBLISH (d0, q2, r0, m31, '/aaa/8', ... (92 bytes))
....
Received PUBLISH (d0, q2, r0, m32, '/aaa/8', ... (92 bytes))

Revision history for this message
Sang kyeong Nam (nskpc1) said : 		#3

Connection failure issue in before comment is not linux issue.

After Result,
I try connect to server from the machine(include receivers) with different id. It's success.
I try connect to server from the machine(include receivers) with same id. It's failure.
I try connect to server from another machine. It's success.

Revision history for this message
Roger Light (roger.light) said : 		#4

I'm just looking at this now. I'm not entirely sure what you mean in your latest comment - can you give me a step by step guide to what you're doing in each of the three cases?

Revision history for this message
Sang kyeong Nam (nskpc1) said : 		#5

1. Execute mosquitto.
2. Receivers connect to server with QoS 2, clean-session is false. And subscribe.
3. Disconnect all receiver. Don't unsubscribe.
4. Sender send 30 messages each receiver.
5. Reconnect all receiver when sending message is completed.

Result.
  38 Receivers fail to connect after test flow 5.
  1860 message receive successful.

after get result
6.1 I try connect to server from the machine(include receivers) with different id. It's success.
6.2 I try connect to server from the machine(include receivers) with same id. It's failure.
6.3 I try connect to server from another machine. It's success.

Revision history for this message
Sang kyeong Nam (nskpc1) said : 		#6

on 6. 2 mosquitto print this message

Client /aaa/2 has exceeded timeout, disconnecting

Revision history for this message
Sang kyeong Nam (nskpc1) said : 		#7

I have a question.

I subscribed twice in keepalive time and send message, receiver receive two message.
Is it true?

Revision history for this message
Sang kyeong Nam (nskpc1) said : 		#8

Bug #741380

This bug affect to my test.
I change code to solution. Then, I solve connection failure issues.

Revision history for this message
Roger Light (roger.light) said : 		#9

So to be clear, this is no longer a problem?

If this is the case, could you mark the question as solved?

Revision history for this message
Sang kyeong Nam (nskpc1) said : 		#10

OK.. But, I have another question.

I will ask you on new thread.