[jira] [Created] (PLC4X-279) nioEventLoopGroup thread proliferation

["Stefano Bossi (Jira)" <ji...@apache.org>]

Stefano Bossi created PLC4X-279:
-----------------------------------

             Summary: nioEventLoopGroup thread proliferation
                 Key: PLC4X-279
                 URL: https://issues.apache.org/jira/browse/PLC4X-279
             Project: Apache PLC4X
          Issue Type: Bug
          Components: Driver-S7
    Affects Versions: 0.8.0
            Reporter: Stefano Bossi
         Attachments: Screenshot 2021-02-01 at 09.19.07.png, Screenshot 2021-02-01 at 09.20.42.png, Threads total 2021-01-31 at 11.34.15.png, Threads total 2021-02-01 at 09.14.02.png

Dear developers,

I did a stability test with the plc4x 0.8.0 and pool2 library.

The test is just reading continuously a couple of DataBlock from a S7 Siemens 1200 PLC.

During the test I have used VisualVM like profiler to inspect the use of the Heap and the threads and I found something strange.

The test runs for a couple of days and I have sampled the threads in a couple of snapshot found that the number of running nioEventLoopGroup thread increase and doesn't seems to be bounded by a limit.

These are the total of the Threads of the applications at 2021-01-31 at 11.34.15; the total number is 246 and the number of nioEventLoopGroup is 192

!Threads total 2021-01-31 at 11.34.15.png|width=568,height=332!

These are the total of the Threads of the applications at 2021-02-01 at 09.14.02; the total number is 358 and the number of nioEventLoopGroup is 305

!Threads total 2021-02-01 at 09.14.02.png|width=549,height=309!

 Form the profiler two of this thread are:

!Screenshot 2021-02-01 at 09.19.07.png|width=530,height=210!

!Screenshot 2021-02-01 at 09.20.42.png|width=539,height=271!  

 

It seems a leakage and seems a dangerous one. 

I think the problem is inside the pool2 library but I am note sure. 

What do you think? 

Regards,

S.

 

 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)