Job Queue through NAV Webservice for mulitple Companies

ara3n

I had posted on mibuso a while back on how to use two NAS instances to run job queue for unlimited amount of companies. Here is the link

viewtopic.php?f=7&t=22633. You had to purchase an additional NAS user license in order to use it. Dynamics NAV 2009 comes with web service, and you can publish any Code unit or Page by default. You will need to purchase one user but for webservice, but it much cheaper. In this blog, I will show you how to setup JOB QUEUE to schedule to run jobs across companies. The advantages of using Web Services is that you can run multiple jobs at the same time for multiple companies. What I’ve experienced with NAV 2009 Web Service is that once you call a web service function, you can disconnect and Service Tier will finish the job and commit the transaction.

In addition you don’t have to modify any processing/posting jobs that have dialog windows displayed. The Service tier simply ignores them. For example you can run adjust cost report without modifying any objects from Job queue. In addition, the solution I’ve created does not modify any existing objects!!!

You can download the fob and read more about it on my blog.
http://mibuso.com/blogs/ara3n/2009/01/2 ... companies/

kine

What to say... =D> =D> =D> 8)

munib00

Top quality.

garak

if i click on the link, the following message is displayed

WordPress database error:
INSERT INTO wp_38_ras_count (id) VALUES (NULL)

WordPress database error:
INSERT INTO wp_38_ras_image (id, createtime, word) VALUES (0, 1233215879, 'w6x27rn')

WordPress database error:
DELETE FROM wp_38_ras_count WHERE id = 0

Waldo

The idea is great! =D>

But I have to think it through though... because when scheduling codeunits through web services ... what about the lack of optimistic concurrency?

ara3n

garak wrote:

if i click on the link, the following message is displayed

WordPress database error:
INSERT INTO wp_38_ras_count (id) VALUES (NULL)

WordPress database error:
INSERT INTO wp_38_ras_image (id, createtime, word) VALUES (0, 1233215879, 'w6x27rn')

WordPress database error:
DELETE FROM wp_38_ras_count WHERE id = 0

I've uploaded in here as well

http://sites.google.com/site/ara3nfiles ... edirects=0

Luc_VanDyck

Blog is working now. I fixed the missing tables.

ara3n

I also submitted it to the download section on mibuso.

ara3n

Waldo wrote:

The idea is great! =D>

But I have to think it through though... because when scheduling codeunits through web services ... what about the lack of optimistic concurrency?

could you clarify?

Waldo

Well, simple ... when two users are running the same webservice, and that web service is a codeunit in NAV ... there is no optimistic concurrency. You have to handle this manually.

kine

It will be handled in same way, as when you run these two codeunits in two different clients. There is nothing specific for WebServices in this... ot it is? :-k

Waldo

yes there is.
You won't get the message "the record has been changed since you last read it from the table...". No version control.

Or at least, this has been told to me...

ara3n

I just tested it and I do get the error.

Another user has modified the record for this Item after you retrieved it from the database.  Enter your changes again in the updated window, or start the interrupted activity again.  Identification fields and values:  No.='1000'

From webservice.

My test codeunit looked like this.

Item.GET('1000');
SLEEP(10000);

Item.VALIDATE(Description,'Bicycle WS');
Item.MODIFY;

So while the WS was waiting I modified the description on the item with another client.

Waldo

Sorry guys, I can't get into too much of detail .. it's just a warning I have been given a couple of times.
I wish a Microsoft dude (freddy? ) would stick his nose into this issue, because also I would like to know the details of it ... .

ara3n

So how do we get his attention? 8)

Waldo

pm?

ara3n

The way I I understood lack of optimistic concurrency was only when you published function that accepted a record. For example.
You publish the customer Page, and then create a new Codeunit with function UpdateCustomerPostingGroup and it takes Customer as parameter. In Web Service Table you publish the codeunit with same Name as Customer. And in Visual Studio when you consume the Customer Page you'll see UpdateCustomerPostingGroup as a function to which you can pass the Current Record you are modifying. And that's where you'll see the issue.

Waldo

They 're not picking up this thread, are they?

DenSter

I think they are reading it, but not answering. Maybe it's a touchy subject? :-k

freddy.dk

No such thing as a touchy subject :?

Anyway - it is correct that there is no optimistic concurrency and there are several ways to cope with this - not one single answer.

I would love to discuss this, understand the issues you run into and discuss various solutions to these (which we then could post here). Any chance that some of you are going to Convergence in New Orleans? - maybe we could setup a discussion on this topic there?

ara3n

ara3n wrote:

I just tested it and I do get the error.

Another user has modified the record for this Item after you retrieved it from the database.  Enter your changes again in the updated window, or start the interrupted activity again.  Identification fields and values:  No.='1000'

From webservice.

My test codeunit looked like this.

Item.GET('1000');
SLEEP(10000);

Item.VALIDATE(Description,'Bicycle WS');
Item.MODIFY;

So while the WS was waiting I modified the description on the item with another client.

If there is no optimistic concurrency? I wouldn't get the above error, and I wouldn't be able to modify the item record. Web service had to lock the item record when it issued a get.

freddy.dk

Ok - you are right, my bad....

We are using optimistic concurrency with no conflict resolution - so if two users modify different fields in the same record, the last user to update gets the error. Web Services here is just like a user.

You can provoke this with 2 clients as well - open 2 clients (Classic or RTC) - open item 1000 in both - modify something in one client and close the card. Modify something in the other client and close the card and you get the error - I guess it has been like that forever.

So the webservice codeunit is the same - the

ITEM.GET

is reading the record and if anything changes the record from that time and until it tries to write the record back - it will fail.

Re-reading the thread - I guess the question is whether or not this applies to codeunits that are published as extention functions to pages - right?
I am pretty sure that it does - meaning that you cannot do like this:

Item_Service service = new Item_Service();
service.UseDefaultCredentials = true;
Item item = service.Read("1000");
service.Test(item.Key);   // <-- This is the function updating the Description
item.Shelf_No = "FF";
service.Update(ref item); // <-- This will fail

You would have to re-read the Item after calling the Test function, since the extension function gets the key and re-reads a representation of the record itself.
That is at least what I know/have tried - let me know whether there are anything I should try to investigate?

ara3n

That's good news, and thank you for the confirmation.

Waldo

yes indeed, this is good news [-o< .

I'm going to be at Convergence ... and i would love to go into this furthermore

elwin68

I have used the installation of the NAS and webservice on our own server last week. Everything worked fine until today.
Nothing is changed at the codeunit RunJob. The only change at the server is a windows update for Internet Explorer 8.

Today the Job Queue Log Entries give me the following error: Http Error 500: Internal Server Error

The event viewer of the server gives me the following error:
Type: Microsoft.Dynamics.Nav.Service.WebServices.ServiceBrokerException
Message: Method "RunJob" is invalid!

Does anyone can give me a clue?

Thanks in advance.

ara3n

check and make sure the you can access the web service through IE and that the runJob is published.

elwin68

I have already done that.
The webservice is running and accessible and RunJob is published.

ara3n

I am guessing you've already restarted the service tier?
Next idea would be to reinstall the server tier.

freddy.dk

I am not sure that I understand what the questions are?

ara3n

hello Freddy. I was asking about the following thread.

viewtopic.php?f=32&t=34672

elwin68

I have restarted the service tier and restarted the computer.
After this I have made a virtual PC with NAV2009 and added the codeunits and the NAS.
The same error occures: Http Error 500: Internal Server Error

The event viewer gives the same error too: Message: Method "RunJob" is invalid!

When have have more time I will try it again. Thanks for your time for now.