Skip to main content


Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Eclipse Projects » 4DIAC - Framework for Distributed Industrial Automation and Control » v1.10.1 Woes(Some recently surfaced issues )
v1.10.1 Woes [message #1801402] Fri, 18 January 2019 10:51 Go to next message
Daniel Ng is currently offline Daniel NgFriend
Messages: 23
Registered: October 2017
Junior Member
Hello guys,

I've been using 4DIAC for quite awhile and recently upgraded to v1.10.1 from v1.9.

I've noticed that after creating composite function blocks, I sometimes get this error "Could not create event connection in CFBERROR" while trying to deploy the FBs. I could narrow down to the problematic FB but I couldn't see visually on the IDE what's causing this error.

Another thing I've noticed is that the monitoring system seems to have been crippled. I could easily monitor many more FBs previously that I can now (which isn't many at all) before the thing sorts of goes blank and i restart it by un-checking and rechecking the "Monitor System".

Any ideas on how I could solve these issues?

Thanks in advance!

Regards,
Daniel
Re: v1.10.1 Woes [message #1801421 is a reply to message #1801402] Fri, 18 January 2019 18:58 Go to previous messageGo to next message
Alois Zoitl is currently offline Alois ZoitlFriend
Messages: 866
Registered: January 2014
Senior Member
Hi,

thanks for bringing this issues to our attention.

Interestingly in a test today I stumbled across the same error message when testing a CFB. In my case the reason was that my 4diac FORTE did not had all the FBs in the composite. Could this be the case for your 4daic FORTE as well?

Unfortunately we have still an issue in detecting missing FB types for CFBs [1].

We did some changes regarding monitoring in 4diac IDE (i.e., harmonizing deployment and monitoring), but I would't have expected to have such an impact. It would be great if we can narrow this issue down. Therefore I would like to ask a few questions here:
- Can you expand on what you mean with ā€˛things sorts of goes blank"?
- Is the issue building up over time or because of the number of watched ports?
- Are you watching on more then one or more devices?
- How many FBs ports are you watching when the issues are starting?
- From your observation do you think its happening on 4diac IDE or 4diac FORTE side?

Especially on FORTE side we have some limitations regarding memory and network that may limit the number of ports that can be watched at the same time.

Cheers,
Alois


[1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=506434
Re: v1.10.1 Woes [message #1801495 is a reply to message #1801421] Mon, 21 January 2019 08:40 Go to previous messageGo to next message
Daniel Ng is currently offline Daniel NgFriend
Messages: 23
Registered: October 2017
Junior Member
Hi Dr Alois,

You were right on the CFB error. I left out on one of the internal FBs in the CMakeLists while doing modifications.

I'm running the 4diac FORTE on four RPI3s that i have connected through Ethernet. While monitoring, I usually would monitor across all devices but not every single FBs in each of these devices. What I meant by things go blank is when the monitoring fails and i see blank yellow highlights on the IDE.

index.php/fa/34696/0/

I'm leaning towards it being an IDE issue but I am able to resume monitoring if I deactivate it and reactivate monitoring on the IDE side without redeployment to FORTE. What I've noticed is that this issue surfaces easily most of the time when I try to watch multiple FBs altogether concurrently when I select all of them together and press watch. The problem doesn't occur when I select FBs one by one to watch, even when I've eventually selected everything for my application. So it seems like its the IDE.

Hope this helps.

Regards,
Daniel
Re: v1.10.1 Woes [message #1801529 is a reply to message #1801495] Mon, 21 January 2019 20:09 Go to previous messageGo to next message
Alois Zoitl is currently offline Alois ZoitlFriend
Messages: 866
Registered: January 2014
Senior Member
Hi Daniel,

great that we could solve your first issue.

The second one is rather strange. I normally do activate watches more then one FB at the same time. The cleared monitoring items has been introduced in 1.10.0 to show that the connection to the device was lost. This can be because of a communication issue or some exception in the monitoring handling code. Do you have some trace outputs in .log file that could help us to narrow it down? Or do you have a recipe that leads always to that behavior allowing us to reproduce it?

Cheers,
Alois
Re: v1.10.1 Woes [message #1801534 is a reply to message #1801529] Tue, 22 January 2019 02:46 Go to previous messageGo to next message
Daniel Ng is currently offline Daniel NgFriend
Messages: 23
Registered: October 2017
Junior Member
HI Dr Alois,

It is fairly easy to reproduce the issue with my application. I've recorded three videos to replicate the problem for you.

https://drive.google.com/open?id=1ThmG4Q7h0IRHWeQMmlx05oA4-q0AOcdy

In V1, I could easily monitor 3 applications that's each on three different RPIs with no issue until the very end where i went into a CFB. In V2, the issue started when i was just monitoring the second application on a second Rpi. Notice there was a lag after pressing watch then monitoring failed. Lastly, in V3, I only watched FBs from one device and went into one of the CFBs to watch that triggered the problem. I then restarted the monitoring on the IDE before trying to watch more FBs on another device which then triggered the problem again.

Let me know how to obtain the .log files and I'll gladly reproduce the problem.

Regards,
Daniel
Re: v1.10.1 Woes [message #1801554 is a reply to message #1801534] Tue, 22 January 2019 13:22 Go to previous messageGo to next message
Alois Zoitl is currently offline Alois ZoitlFriend
Messages: 866
Registered: January 2014
Senior Member
Hi,

thanks for the videos. I'm currently traveling. I'll try to look as soon as possible on them.

You can find the .log file in the .metadata directory in your workspace.

Cheers,
Alois
Re: v1.10.1 Woes [message #1801581 is a reply to message #1801554] Tue, 22 January 2019 21:26 Go to previous messageGo to next message
Alois Zoitl is currently offline Alois ZoitlFriend
Messages: 866
Registered: January 2014
Senior Member
I found some time today in the train to look on your videos and your code. After the first I thought the cause is that you are monitoring inside of CFBs. But the other videos showed my that this is not the case. I think that it could be a race condition in the device communication code. This would be quite quick to fix. However with my tests I was not able to reproduce it. Therefore getting some outputs from your logs could really help us.

Cheers,
Alois

P.S.: have you seen the news post on the improvements in deployment, i.e., you can now deploy also from the system perspective. I think for your tests this could save you quite some time.
Re: v1.10.1 Woes [message #1801896 is a reply to message #1801581] Tue, 29 January 2019 10:52 Go to previous messageGo to next message
Alois Zoitl is currently offline Alois ZoitlFriend
Messages: 866
Registered: January 2014
Senior Member
Today we merged a first fix for what could be a cause of your problem [1]. You can try it by building 4diac IDE by yourself. Or wait for the upcoming 1.10.2 maintenance release. Which should come in the next days.

Cheers,
Alois

[1] https://bugs.eclipse.org/bugs/show_bug.cgi?id=543754
Re: v1.10.1 Woes [message #1802027 is a reply to message #1801896] Thu, 31 January 2019 07:04 Go to previous messageGo to next message
Daniel Ng is currently offline Daniel NgFriend
Messages: 23
Registered: October 2017
Junior Member
Hi Dr Alois,

I got sometime today to capture the log messages for you as attached.

Thanks for letting me know of the new deployment method, I wasn't aware of it.

Regards,
Daniel
  • Attachment: log.txt
    (Size: 7.49KB, Downloaded 12 times)
Re: v1.10.1 Woes [message #1802072 is a reply to message #1802027] Thu, 31 January 2019 17:25 Go to previous message
Alois Zoitl is currently offline Alois ZoitlFriend
Messages: 866
Registered: January 2014
Senior Member
Hi Daniel,

thanks for the log. I think what is shown there could really be the concurrency problem I fixed in Bug 543754 as described in my last message. So if you would like to build 4diac IDE from source you can try the latest version from the 1.10.x commit or wait for the 1.10.2 release.

Alois
Previous Topic:Encapsulating listen sockets in Function Blocks
Next Topic:Issues compiling Forte with OPC UA module included in Windows
Goto Forum:
  


Current Time: Wed Aug 21 18:21:29 GMT 2019

Powered by FUDForum. Page generated in 0.02439 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software

Back to the top