DSC: tormented by error 1179

[David Moerman]

I'm getting error 1179 (Access mode not supported for operation) on an inconsistent basis.  It is coming from a Datasocket Write function which I'm using to update shared variables.  The datasocket SV references are opened in read/write mode (but without using the "?sync=true" suffix which seems to make no difference to the problem anyway).  Basically how it goes is this:  I deploy a small library of SVs (about 20 of them) and then immediately open datasocket references to each of the SV's, then a little later try to write to them, which is when I get the error.  It is very inconsistent:  sometimes the procedure works without a flaw, but a little later I run it again and get error 1179 at the Datasocket Write function.  Also, not all of the variables will give the error; some will and some won't, and even that won't stay consistent on the next problematic run.  However, on a given run once the variables give an error, they continue to error until I close the process and try it all again.

Using LV DSC 8.5, and I'm logging each of these SVs to a remote Citadel database, on another computer on the local subnet.  Does anyone have any ideas?

Thanks,
--David Moerman

[David Moerman]

Some additional info:
1.  Once a deployed process gives me this error, if I stop the process and re-deploy it, then it will ALWAYS continue to give the error
2.  Turning off DSC logging altogether this has no effect:  the error persists
3.  Restarting LabVIEW has no effect
4.  If I rename the library and deploy it, I usually DON'T get the error
5.  If I rename the libary back to the first name and deploy, I usually DON'T get the error

--David Moerman

[Eriquito]

Hi David,

That is an unusual error.  I looked through our archives and didn't see any reports of any similar issues.  Have you tried accessing the shared variables directly using shared variables nodes?  My guess is, this would resolve your problem.

If that solution doesn't work for you, perhaps you could post your project files so we can attempt to reproduce the problem here?

[David Moerman]

Firstly, scratch point #5 in my previous message.

More importantly, I've discovered something repeatable that is likely part of the same issue.  However, this procedure produces error 1179 in Datasocket Open, whereas the previous error I mentioned manages to open the SV successfully and then later produces 1179 during Datasocket Write.

It seems to have something to do with Variable Manager (which I often use to monitor the processes during my program), and an interplay with the "Single Writer" option in SV properties.  I've included a zipped project folder for demonstration.

1.  Open Variable Manager.

2.

[David Moerman]

Sorry, my spacebar landed on the "submit post" button

2.  Deploy Var_Lib.lvlib.  Notice only a single shared variable using the "Single Writer" option.

3.  Open and run "Using_Datasocket.vi".  Notice no problems.

4.  Refresh the Variable Manager and add "output1_SV" to the watch list

5.  Run the VI again.  Error 1179.

This problem does not occur if the variable is not part of the watch list, does not occur if the Variable Manager remains closed, and does not occur if "Single Writer" remains off.  However, once it occurs for the process, the only way to get rid of it is to shut down the SV engine and restart it.  Simply restarting the process doesn't seem to work.

--Dave

Message Edited by David Moerman on 05-19-2008 09:05 PM

Message Edited by David Moerman on 05-19-2008 09:06 PM

[David Moerman]

OK, another clarification:

Restarting the SV engine is NOT the only way.  I just realized that to get rid of the error one may remove the variable from the watch list OR shut down the Variable Manager. 

This may explain my difficulties.  I would check various variables in the watch list at different times, some of them may have had the "Single Writer" option in use.

--Dave

[Eriquito]

Hi David,

I have reproduced the 1179 error following the procedure you outlined.  I'm not entirely sure what's going on here, but I'm looking into it.

However, for me (in 8.5.1), simply removing the variable from the watch list and running the program again makes the error code go away.  Is this not the behavior you are experiencing?

[David Moerman]

Yes, that right -- it turns out that is the easiet way to get rid of the error.  So hopefully, that is the main issue.  I will post back if there is anything else I find that influences it.

Thanks,
--Dave

[Eriquito]

At first glance, this appears to be a bug.  This was reported to R&D (# 113075) for further investigation.

FYI, I noticed that if you add the variable to the watch list, CLOSE the variable manager, and the VI runs without error.  While vi is running, you can open the variable manager and it still works fine.