Re : Re: Re : Error Code -1950679023

[SGCLuis]

Hi, I´m having a problem reading and writing shared variables in Labview. I´ve created a project where I have a real time target (cFP2100) and a PC. In my RT target I´ve deployed a software that´s reading and writing some shared variables. I´ve also deployed in my RT target the variables library. In the software that I´ve deployed in my RT target I have two loop cycles running in parallel. One of the cycles is running with a 25ms delay. The second cycle is running with a 200ms delay. It is in this 2^nd cycle that I´m reading and writing my shared variables. When I deploy my software I first build it, then set as startup and then run as startup. Sometimes I get the following error message: LabVIEW:  Failed to load shared library  sysstatepublisher.dll on RT target device. This error disappears when I repeat the procedure for a second time (bulid, set as startup, run as startup). I then see that the software is running ok in my cFP target, because I´ve placed a heart beat in both loop cycles that makes two leds of the cFP´s front panel blink from time to time. I´ve also introduced a 20s delay in the beginning of the software, because I´ve read that after a reboot, the shared variable engine needs sometime to initialize properly. My problems beggin when I try to read and write the shared variables from my PC. I´ve created I small VI which I build and run it as a standalone application. This is nothing more than a simple loop cycle with a delay of 200ms to read and write to the shared variables in my cFP2100. When I run this application I first get the warning 1950679027 and a few second after I get the error 1950679023. I only run this application after I have the two leds blinking in my cFP. This is to make sure that the reboot process is complete.  If I try to stop and run a few times this application, the 2^nd cycle in my cFP software (the one reading and writing my shared variables) stops. I know this, because I see that the correspondent led in the cFP2100 front panel stops to blink. The led regarding to the first cycle continues to blink… I´m a bit lost here because from time to time I seem to get things working, but if stop everything and start again, I can´t make it work.I´ve checked in MAX and I see that I have installed in my cFP2100 the Network Variable Engine 1.5.0 and the Variable client support for Labview 1.5.0. I´ve checked in the Distributed System Manager and I see that the variables are deployed in the RT target. I´ve tried to disconnect my firewall and anti-virus. The shared variables are set in their properties to RT FIFO. The only thing that is strange is that I´ve read from one of your forum´s post, that I should have the NI Shared Variable Engine entry in add/remove program option found in the Windows Control Panel, under the National instruments software. I´ve checked and I don´t have it, but if this is the problem, how can I from time to time be able to get the system working?My system is as follows:·         Window Vista Home premium, service pack 1·         Labview 8.6 full development system with application builder and realtime module.·         Fieldpoint 6.0.2·         cFP2100 and cFP1800 expansion bank (both network connected to my PC) Can anyone help? Thanks in advanceBest regards Luís

[ajangot]

Hi,

Thanks to use National Instruments forum.

I see few solutions to your problem:

First, upgrade your NI-Fieldpoint 6.0.2 to 6.0.4. You can found here: NI-FieldPoint 6.0.4 - PharLap, VxWorks, Windows 2000/Vista x64/Vista x86/XP 

Your error 1950679023 is maybe provok by the windows firewall Using Lookout, LabVIEW DSC and Network-Published Shared Variables with Firewalls

I advise you to read this: Troubleshooting Network-Published Shared Variables

Regards,

[SGCLuis]

Dear Aurelien,

First of all thank you for your help and sorry for my delayed reply.

After I´ve posted my question,  I tried to use another CPU (cFP2200). Although I still got the same error I´ve described while deploying the application (LabVIEW:  Failed to load shared library  sysstatepublisher.dll on RT target device) things looked a lot better after a successful deployment. I can see the variables being updated in my SCADA VI running in my computer and I don´t get error 1950679023. I´ve already updated my fieldpoint version to 6.0.4, but I still get that error from time to time while deploying. For now I´ll be using this new CPU. Later I will format the cFP2100 CPU and install all the software again. As for the error I get while deploying, can you give some more detail about it? Since I get the same error using 2 different CPUs, I suspect that the problem is not in the CPU but in the Labview installation in my computer. Also can you please inform if cFP2100 has enough memory resources to host the shared variables library?

Best Regards

Luís

[ajangot]

Hi,

cFP2100 and cFP2200 have enough memory and CPU to run sysstatepublisher.dll. This DLL allow to see memory and CPU in real-time. I don't know if you realy need of the feature?

I think that you have to reinstall your NI-DAQmx: NI-DAQmx Download

Regards,

[Riconquistiamola]

Luis,

First and foremost, please do not reinstall DAQmx, as it has nothing to do with your problems (and isn't even available for install on FieldPoint controllers).

I must admit your first post is very difficult to understand, so I have a few bullet points to cover.

System Setup: cFP-2100 (cFP-2200 also tested) running startup RTEXE. Shared variables are hosted on the target. SCADA reading shared variables on host.

Issues:

• Two shared variable errors (NI-PSP has not connected to server yet, and connection to server was disconnected) when running SCADA after target RTEXE has initialized (confirmed by 2 LEDs).
• Failure to deploy RTEXE on first try because sysstatepublisher.dll fails to load.

Questions:

• Is this an accurate description of your setup?
• Are the shared variables network-published? Is the RT FIFO enabled? The Shared Variable Engine is necessary if they are network-published and hosted on the target. I can't explain how this worked without it.
• Does your RTEXE use CPU/Memory VIs in the RT palette? These are the only VIs dependent upon sysstatepublisher.dll (System State Publisher, the DLL allowing NI-DSM and other system state tools to work).

Aurelien, please continue working with Luis after this information is gathered. It would be nice to reproduce this behaviour internally for further investigation.

Cheers.

[SGCLuis]

Dear Michael, Thanks for your reply to my post. Before I answer your questions, I must inform that things are working much better now. This was achieved only after I´ve changed the CPU from cFP2100 to a cFP2200. Before this, and even after I´ve upgraded my Labview to 8.6.1 and my Fieldpoint to 6.0.4, things simply didn’t work with cFP2100. My current system is Labview 8.6.1, Filedpoint 6.0.4 and cFP2200. Regarding to your questions, and going back to the system configuration I had when I first sent my post, my feedback is as follows: 1 - Your description is accurate.2 - My shared variables are network published and have always been. RT FIFO was enabled at the time. It´s desabled right now. As mentioned, I had checked that Network Variable Engine 1.5.0 and Variable client support for Labview 1.5.0 were installed in the RT target. I also mentioned that I couldn´t find NI Shared Variable Engine entry in the ADD/REMOVE PROGRAM in the windows control panel. I still can´t find it and things are working much better. Maybe it´s under another name...

3 – I´m not using and wasn´t using at the time, any VIs from the RT palette. If you noticed above, I said that things are working much better now. The reason why it’s better and not perfect is because I still get that sysstatepublisher.dll error from time to time.

I hope this information is clear and helpful. Let me know If you need me to run more tests. I want to understand what the problem was.

For now I will keep my current system configuration which seems to be stable.

Cheers

Luís 

[ajangot]

Hi,

Indeed, it's not necessary to reinstall daqmx, sorry for this misunderstanding.

I have few questions about your configuration:

1. It's better with cFP2200 but what have you change between your cFP2100 and your cFP2200?
2. What is your network configuration? Your PC and your fieldpoint are in direct connection or throught a routeur? What is show in MAX>>cFP target>>advance network
3. Can you send your projet or a simple exemple with your error?
4. In the Distributed system manager, you say that you see your variable but are they update?

Regards,

[SGCLuis]

Dear Aurelien,I`m very sorry for my late answer. I´ve been busy developing my projectAfter I´ve replaced the cFP2100 with a cFP2200, things started to work immediately better. I didn’t do any more changes. The code is the same and apart from the new CPU, all other hardware is the same, including network hardware (all I have is a simple 8 port hub between my PC and the RT target).  I think that replacing the CPU was the answer to my problems. As I mentioned in my previous post I still get the sysstatepublisher.dll error from time to time, but for now I can live with it. Later I will format the cFP2100 and install all the software again, and then run more tests on it.For now I will keep my current configuration, and continue to develop my project. I´ll get to you when I have any more problems (hope not…) and to let you know about the tests in the reformatted cFP2100.Thank you and Michael for your time and help.Best regards Luís

[ajangot]

Hi,

Thank you for this feed-back, I waiting for your result of tests.

Regards,

[ulises.uribe]

Equipement setup: cFP2100, Fieldpoint 6.0.5, LabVIEW 2009

Hi everyone I`m also having this problem but a remarcable sympthom must be clarified: when I open the NI Distributed system Manager to monitor  the  CPU load it displays 100 % from the very beggining. I know that the code couldn`t be because I`m running a simple while loop with an indicator. Any suggestions?