Data socket problems with LVRT 8.2

[chatonda]

 I am in the process of migrating to LV 8.2 from LV7.1.1. My system has 18 RT targets which send data to a Host PC through data socket (DS). The Host PC runs LV on Windows XP Pro.  With LV7.1.1 on both host and RT targets DS works perfectly. The RT targets update the host every 500 ms.

I've taken a three-step process in the migration. First tests were done with the host migrated to LV8.2 while keeping the RTs at LV7.1.1. Result: the system worked perfect. Then I changed one RT to 8.2 keeping the rest the rest at LV7.1.1. Result: DS timed out before all the data could be sent to Host. Finally, all RTs were migrated to LVRT 8.2. Result: Only about 5 out of the 18 RTs report to the Host before DS times out. However, if only one RT target is used, i.e., the other 17 RTs are diasbled, communication is very fast. Is there a special setting that I need to make? Btw, the roughly 5 RTs that report vary randomly from test to test, i.e., there is no bias towards any particular RTs.

I am using the same source code developed under LV7.1.1, so the only changes made are the ones that take place when one opens LV7.1.1 code in LV8.2. Some VIs get converted in the process.

Any ideas?

Chatonda Mtika
Algis Corp
Vancouver, Canada

[Nadim R]

Hi Chatonda,

Before we dive into this issue, I have a few questions:

1)  You mentioned that 5 out of the 18 RTs report data to the host before the time out.  You also mention that if you run one RT, it works perfectly.  My question then becomes, is it the same RTs that timeout? even when connected individually? 
2)  When you migrated to LV RT 8.2 on the targets, did you upgrade/change any software on the targets?
3) What happens when you connect one RT target and add targets and test sequentially?
4)  What version of the datasocket for LabVIEW RT do you have on your targets?

Regards,

Nadim
Applications Engineering
National Instruments

[chatonda]

Hi Nadim, Question 1: it doesn't matter which RT you use; it will work perfectly as long as it is the only one connected. Question 2: I installed LV RT 8.2 plus all the associated software that comes with LVRT 8.2 on the RTs. Question 3: I did not try this; it would be very tedious in my setup. Question 4: Datasocket 4.2 for LVRT 7.1.1 and 4.4 for LVRT 8.2 Thanks, Chatonda

[Nadim R]

Hi Chatonda,

We have noticed that you have opened up a discussion forum as well as a phone support request.  We would like to keep things simple and refer all support questions to the phone, rather than to both the phone and the discussion forum.  Please reference your SR number when calling in for support and we will be glad to assist you.

Regards,

Nadim
Applications Engineering
National Instruments

[Ben]

"We would like to keep things simple and refer all support questions to the phone, rather than to both the phone and the discussion forum. '

There are some of us that are interested in the outcome of this thread.

Please post a summary of the -offline daventures and let us know what the fixes are.

It's only a mtter of time until I will be walking across this same UPGRADE bridge and I'd like to know what I am in for.

Ben

[chatonda]

Ben,

I must say I am very surprised by Nadim's email. His point, which I must say I don't agree with, was made quite clear to me yesterday during a conference call between my team and the NI team comprising Chris, Nadim, and Avinash. So I don't know what purpose this morning's email is supposed to serve. Be that as it may, I think NI has gotten it wrong here. I do not think it should be one or the other. The discussion forum is a much wider forum comprised of people with varying experiences - not just NI employees. That is precisely why I posted my problem to the forum: to tap into the vast experiences of the discussants. But when I did not get any responses after I had posted answers to Nadim's questions, I decided to make a formal service request to NI, all perfectly legal. So I was really surprised when I was being blamed for tying up NI resources on one problem. I believe the resource allocation issue should be NI's to solve, not mine. I have no way of knowing that NI had formally assigned somebody to my problem. What if nobody from NI responds, am I allowed to make a service request to NI? My feeling is that NI's logic is flawed here, assuming Nadim's views are NI's views.

Thanks,

Chatonda Mtika
Algis Corporation
Vancouver, Canada

[Nadim R]

Hi Chatonda,

Let me make it clear that it is perfectly legal to post a discussion forum as well as open a service request on the same issue.  I posted the response earlier this morning to make it clear to others reading this discussion forum that we are taking care of this issue on the phone.  I will by the way, post results on this discussion forum once the issue is resolved.  I also posted this response to make it clear that we need one directive - through the phone, rather than have me post responses to the discussion forum and have other AEs work on the issue on the phone.  It will be quite confusing to do so.  I am not saying that other people in the community should not respond here to the discussion forum.  If they have any sort of information that may pertain to this issue, the better.  It's more of a a matter of directive.  I hope I answered your concerns.


Regards,

Nadim
Applications Engineering
National Instruments

[Ben]

Hi Chatonda and Nadim,

I have a LV 6.1 RT app running that is waiting for funding to get upgraded. What ever you find may help me when that upgrade hits my desk.

Nadim,

Please be patient with us! Many of us operate on a time scale such that a fix is expected in days as opposed to weeks. When we don't get updates, we get desperate, welll at least I do.

Take care,

Ben

[ashm01]

hello All,

Considering I am running into a similar communications issue on a particular controller , I am very much interested in the final outcome / resolution of this problem. Anyone kind enough to fill us in ....

I simply can't believe a thread is left hanging without an outcome posted.

Ashm01

[Nadim R]

Hi,

We were not able to replicate the issue found in this discussion forum here at NI Austin-  however, the solution that worked in this case was to change from dstp protocol to psp.

Regards,

Nadim
Applications Engineering
National Instruments