Pitchblack

Manually installing the SOURCECODE data source in SSRS2008 and Windows server 2008

NeilM — Fri, 07 May 2010 14:59:00 GMT

As you probably know, K2 4.5 allows you to run the K2 reports without requiring a SSRS installation. This is all good if you are only using the standard reports, but what happens if you have custom reports that require that the SOURCECODE data source is installed on your SSRS server?

well, while you CAN install the Reporting Services component on the SSRS server using the K2 installer, this may have unintended side effects, since it will also deploy the standard K2 reports to your SSRS server, and (as far as I can tell) will reconfig workspace to point to the reports deployed on SSRS. If you can live with this, then good. if not, then read on....

So, here's what you have to do to manually register the SOURCECODE data provider in SSRS 2008. (Note: this will probably work for SSRS 2005 as well, and Server 2003. I just completed these steps on a SSRS2008 and Server 2008 environment.)

WARNING: because you have manually installed the data source in SSRS, when you upgrade K2 to a newer version, you will probably have to update these assemblies and potentially the assembly registration manually as well. Don't say I didn't warn you! :)

OK, first, you have to copy some assemblies from your K2 server. Head over to the K2 install directory (by default something like C:\program files\K2 BlackPearl\Bin) and copy the following assemblies to a temporary location:

- SourceCode.Data.SmartObjectsClient.dll
- SourceCode.Categories.Client.dll
- SourceCode.HostClientAPI.dll (Please see 64-bit note below!)
- SourceCode.SmartObjects.Client.dll
- SourceCode.SmartObjects.Management.dll

Note for 64-bit operating systems: The SourceCode.HostClientAPI.dll is compiled to target a specific CPU platform. If you are running a 64-bit operating system in your SSRS environment, you HAVE to copy the correct HostClientAPI.dll file FROM THE GAC on the K2 server. It will be located in a directory like C:\Windows\Assembly\[GAC_64\AMD_64]\SourceCode.HostClientAPI\[4.0.0.0_16a2c5aaaa1b130d]

If you copy the wrong HostClientAPI.dll, you will get an error like 'attempted to load a program with an incorrect format'

OK, the next step is to register these assemblies in the GAC of the SSRS server. On Windows 2003, you can probably just drag+drop the files into the Windows\Assembly folder. On Windows server 2008, it ain't that easy.

If you have the gacutil.exe utility installed on your Windows server, you can continue to adding the assemblies to the GAC. If you do not, you will need to download the windows 7 SDK from Microsoft in order to get GACUtil installed (you can probably hack it and try to install it manually, but at this stage I was done with hacking installs, didn't need the extra overhead of hacking a MS install as well). You can find the Windows 7 SDK installer from here: http://www.microsoft.com/downloads/details.aspx?FamilyID=c17ba869-9671-4330-a63e-1fd44e0e2505&displaylang=en

OK, so now you have gacutil.exe installed on your SSRS server. Next, open the SDK command prompt or a windows command prompt (NOTE: use the 'Run as Administrator' option to launch the command line, otherwise...yup, you guessed it - it ain't gonna work) and execute Gacutil to register each of the copied assemblies in the SSRS server's GAC. For reference, the commands will be:

- gacutil -i SourceCode.Data.SmartObjectsClient.dll
- gacutil -i SourceCode.Categories.Client.dll
- gacutil -i SourceCode.HostClientAPI.dll
- gacutil -i SourceCode.SmartObjects.Client.dll
- gacutil -i SourceCode.SmartObjects.Management.dll

Whew, OK. Hang in there, still a few more things to do.

next up, we need to register the SourceCode data source in SSRS. Head to your SSRS install directory (this will depend on your environment, but it will be something like C:\Program Files\Microsoft SQL Server\MSRS10.MSSQLSERVER\Reporting Services).

Locate the file rsreportserver.config and make a backup of it. Next, edit the file with a text editor and add the SOURCECODE data source. You will add the following line in the <Extensions><Data> section of this config file.

<Extension Name="SOURCECODE" Type="SourceCode.Data.SmartObjectsClient.SOConnection,SourceCode.Data.SmartObjectsClient, Version=4.0.0.0, Culture=neutral, PublicKeyToken=16a2c5aaaa1b130d" />

That should be it. Restart your Reporting Services service, perform an IISReset, and then browse to SSRS and attempt to register a new data source using the SOURCECODE data provider. Provide a relevant connection string for the provider, and you're good to go.

(At this stage, you may be thinking: what on earth is a 'relevant connection string?!' here are samples - you'll need to modify them for your environment, of course

Integrated=True;IsPrimaryLogin=True;Authenticate=True;EncryptedPassword=False;Host=blackpearl;Port=5555;UserID=K2Student;WindowsDomain=Denallix;Password=K2pass!;SecurityLabelName=K2

Integrated=True;Authenticate=True;Host=blackpearl;Port=5555

Your next step will probably be to resolve the Kerberos authentication issues that will crop up. If your SSRS server and K2 server are on different machines, you will need to set up kerberos delegation to allow SSRS to impersonate users when connecting to SSRS. Or, alternatively, you can edit the connection string of the SOURCECODE data provider to use specific user credentials. if you can live with the security hole this represents.

Hope that helped!

Reporting Services Fails to connect to K2 Data Source

Murphy — Wed, 02 Dec 2009 14:30:00 GMT

In the world of reporting services there are many things that can go wrong. I was dealing with the following error when trying to run a report which was using a K2 data source.

This manifested itself as:

Cannot create a connection to data source ‘BLACKPEARL’.

A connection attempt failed because the connected party did not properly respond after a period of time.

Troubleshooting:

A very useful tool here is to try and telnet to the k2 server from the machine that cannot connect

1) Run command prompt
2) telnet k2servername 5555

If you cannot connect then it is more underlying windows com problem (possibly a firewall) and its not even worth looking at ssrs until this is resolved.

If telnet is not installed then you can install it by following details on this link

http://technet.microsoft.com/en-us/library/cc771275(WS.10).aspx

Resolution:

In my case telnet showed that I could not connect, therefore I looked for any firewall that were in the way and found that Windows Firewall was turned on on the K2 Server.

Add an Exception rule to allow port 5555 (and port 5252 for other operations) as these are the ports K2 uses for communications (unless you specified different ports in the install).

Unable to Connect to SQL Server (Error is 18456, Severity: 14, State: 11)

Murphy — Wed, 02 Dec 2009 14:18:08 GMT

I have recently spent some time looking into an issue where my K2Workspace account was not able to connect to SQL Server.

This manifested itself as:

In the workspace the following error:

- Membership could not be verified

In SQL Management Logs as:

- Error is 18456, Severity: 14, State: 11

Resolution:

To resolve this issue we first looked to validate that connect rights were given to

The account itself
The roles the users were members Of

Finally it turned out that the public role did not have Connect rights for TCP communication as shown below

I hope this helps someone

Complex Preceding rule: No Code

Willem — Tue, 24 Nov 2009 12:10:41 GMT

Business Requirement

A typical business process that may branch to several parallel activities

On completion of the parallel activities, all lines should join up to continue processing with a single instance

Parallel activities may vary and are determined by runtime business rules (Scenario 1 & 2 following)

Scenario 1

Scenario 2

Technical challenges

· Variable parallel activities. In this business case one (1) or more activities may execute at the same time, and is only determined during runtime execution and business logic validation.

· No code. The process is being developed by a ‘empowered’ business user who understands the business process and is able to create workflows, infopath forms and SmartObjects.

Activity Statuses

· To achieve the requirement of knowing when various independent activities completed a SmartObject is required per logical grouping

· Each activity executed in parallel is represented by a property in the SmartObject

· A single row in the SmartObject represents an instance of the process being executed

· All property values for the row being executed needs to be ‘0’, indicating that there are no activities being processed

K2 Process Design

· The process design at first glance may appear clunky but is not. This process design caters for all of the destination types with regards to slots: ‘Plan just once’, ‘Create 1 slot’ and ‘Create a slot for each destination’

How it works: [Create Status Instance]

· This activity is the activity that creates a single record in the SmartObject.

· For each activity the initial status is set to ‘0’. This is required since some activities may not be executed and will as a result not be able to set a completed (‘0’) status flag.

· The ‘itemId’ returned will be used to update all subsequent calls to the DB

· Once this activity finished processing the ‘Preceding Rule’ Smartobject record will look like this:

activityONEstatus activityTWOstatus activityTHREEstatus activityFOURstatus

0 0 0 0

How it works: [I am busy ...]

· These steps indicate per activity whether or not a specific activity received scope to be executed

· Each activity invokes the ‘Preceding Rule’ Smartobject and sets the value for the related column to ‘1’, this indicates that the activity is ‘busy’

· Why the extra activity?

Performance – in most cases, the activity behaviour with regards to destinations creates a single slot with multiple destinations and ‘Plan-just-once’ is not enabled. If we were to add this event in the event with multiple destinations, a number of unnecessary updates will be executed against the smartobject – one for each destination.

How it works: [I am done ...]

· Once an activity (ONE, TWO, THREE or FOUR) has completed, it has to indicate that by setting the related column value in the ‘Preceding Rule’ SmartObject to’0’.

· Why the extra activity?

Performance – same applies as per [I am busy ...]

Incorrect status value – if the event in this activity was added to the activity (ONE, TWO, THREE or FOUR) with multiple destinations multiple slots, the first person to complete his/her work would update the item status to complete, then when the ‘Continue Processing’ activity preceding rule executes it would evaluate to ‘true’ where in fact it should be false.

How it works: [Continue processing]

· For each line connected to this activity, the preceding rule will evaluate all column values for this instance. Since we already set the columns values for activities to ‘complete’ even though those activities never got scope, stock standard K2 blackpearl functionality could be used to resolve the preceding rule.

· The configured preceding rule:

Sample illustrated

[Create Status Instance] - PrecedingRule SmartObject data

activityONEstatus activityTWOstatus activityTHREEstatus activityFOURstatus

0 0 0 0

[I am busy ...]

PrecedingRule SmartObject data each activity sets the relevant SmartObject property

activityONEstatus activityTWOstatus activityTHREEstatus activityFOURstatus

1 0 1 0

[I am done ...]

· For each of the activities ONE... to FOUR, once they finish execution they set their underlying execution status to 0

· When THREE finishes before ONE the status table will look like this:

activityONEstatus activityTWOstatus activityTHREEstatus activityFOURstatus

1 0 0 0

[Continue processing]

Only when ONE finishes does the preceding rule in ‘Continue Processing evaluate to true

activityONEstatus activityTWOstatus activityTHREEstatus activityFOURstatus

1 0 0 0

Generic Send Mail K2 Exception Handler (Send mail on process error/exception)

hoshy — Tue, 24 Nov 2009 08:26:00 GMT

Hi,

For the last couple of years I've been recommending this as, if not best practice, then minimum practice - i.e. the very least you should be doing when developing processes.

I realise that many out there are building no-code solutions and thus don't / can't benefit from the above advice. Just recently on an internal K2 mailing list one of the guys sent around a useful copy&paste block of code to safely attempt sending an email if a process falls to error state. I can't take credit for this but I wanted to post it here so that all K2 folks who both do and don't do code can benefit.

The ideal location for this code is in the process level exception handler - please see the screen shots below:

First access the process level exception handler wizard:

Then enable the Exception handler, tell K2 to log any unhandled exception and put the process instance in to error state - tick all boxes, then access the code view by clicking View Code:

Expand theExecuteCode section:

Remove the highlighted code:

And paste in the code from the bottom of this blog post:

Make sure you update the email addresses highlighted in red.

Full credit goes to Eugene Jones for the code for this.

Ashley.

Code:

            const string fromAddress = "Administrator@k2demo.local";
            const string toAddress = "BPUser@k2demo.local";

            bool _bFlag = true;
            string _sMessage = string.Empty;
            Exception ex = new Exception("K2Exception");
            object _exContext = null;
            System.Web.Mail.MailMessage _expMail = null;

            try
            {

                K2.AddToErrorLog = K2.Configuration.IsErrorLog;

                K2.AddToServerLog = K2.Configuration.IsServerLog;

                System.Web.Mail.SmtpMail.SmtpServer = K2.StringTable["Mail Server"];

                _expMail = new System.Web.Mail.MailMessage();

                _expMail.From = fromAddress;

                _expMail.To = toAddress;

                // Get the process details

                _expMail.Subject = "Process " + K2.ProcessInstance.Folio + " has gone into error state";

                _sMessage = "Process Folio: " + K2.ProcessInstance.Folio + System.Environment.NewLine + System.Environment.NewLine;

                _sMessage += "Process ID: " + K2.ProcessInstance.ID.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                _sMessage += "Process Fullname: " + K2.ProcessInstance.Process.FullName + System.Environment.NewLine + System.Environment.NewLine;

                _sMessage += "Process Start Date: " + K2.ProcessInstance.StartDate.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                _sMessage += "Process Originator: " + K2.ProcessInstance.Originator.Name + System.Environment.NewLine + System.Environment.NewLine;

                _sMessage += "Process Data: " + System.Environment.NewLine + System.Environment.NewLine;

                foreach (SourceCode.KO.DataField _data in K2.ProcessInstance.DataFields)
                {

                    _sMessage += "Field name: " + _data.Name + "\t" + "Field type: " + _data.FieldType.ToString() + "\t" + "Field value: " + _data.Value.ToString() + System.Environment.NewLine;

                }

                _sMessage += System.Environment.NewLine + System.Environment.NewLine;

                // Get Context and details to the component that threw the exception

                switch (K2.ContextType)
                {
                    case SourceCode.KO.ContextType.ClientEvent:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.ClientEventContext)_exContext).ActivityInstanceDestination.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Event Name: " + ((SourceCode.KO.ClientEventContext)_exContext).Event.Name + System.Environment.NewLine;

                            break;
                        }

                    case SourceCode.KO.ContextType.DestinationRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.DestinationRuleContext)_exContext).ActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Destinations: ";

                            foreach (SourceCode.KO.Destination _dest in ((SourceCode.KO.DestinationRuleContext)_exContext).Destinations)
                            {

                                _sMessage += _dest.Name + " [" + _dest.Type.ToString() + "], ";

                            }

                            _sMessage = _sMessage.Substring(0, _sMessage.Length - 2) + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Slot data: ";

                            foreach (string _slotdata in ((SourceCode.KO.DestinationRuleContext)_exContext).SlotInstanceData)
                            {
                                _sMessage += _slotdata + ", ";
                            }

                            _sMessage = _sMessage.Substring(0, _sMessage.Length - 2) + System.Environment.NewLine + System.Environment.NewLine;
                            _sMessage += "Dynamic Queues: " + ((SourceCode.KO.DestinationRuleContext)_exContext).DynamicQueues.ToString() + System.Environment.NewLine + System.Environment.NewLine;
                            _sMessage += "Resolve Queues To Users: " + ((SourceCode.KO.DestinationRuleContext)_exContext).ResolveQueuesToUsers.ToString() + System.Environment.NewLine + System.Environment.NewLine;
                            _sMessage += "Single Instance: " + ((SourceCode.KO.DestinationRuleContext)_exContext).SingleInstance.ToString() + System.Environment.NewLine + System.Environment.NewLine;
                            _sMessage += "Number of slots: " + ((SourceCode.KO.DestinationRuleContext)_exContext).Slots.ToString() + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.EscalationAction:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.EscalationActionContext)_exContext).ActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Redirect: " + ((SourceCode.KO.EscalationActionContext)_exContext).Redirect.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.EscalationRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.EscalationRuleContext)_exContext).ActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.EventEscalationAction:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.EventEscalationActionContext)_exContext).EventInstance.ActivityInstanceDestination.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Event Name: " + ((SourceCode.KO.EventEscalationActionContext)_exContext).EventInstance.Event.Name + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.EventEscalationRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.EventEscalationRuleContext)_exContext).EventInstance.ActivityInstanceDestination.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Event Name: " + ((SourceCode.KO.EventEscalationRuleContext)_exContext).EventInstance.Event.Name + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.EventSucceedingRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.EventSucceedingRuleContext)_exContext).EventInstance.ActivityInstanceDestination.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Event Name: " + ((SourceCode.KO.EventSucceedingRuleContext)_exContext).EventInstance.Event.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Succeeding Rule: " + ((SourceCode.KO.EventSucceedingRuleContext)_exContext).SucceedingRule.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.IPCEvent:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.IPCEventContext)_exContext).ActivityInstanceDestination.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Event Name: " + ((SourceCode.KO.IPCEventContext)_exContext).Event.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "IPC Process: " + ((SourceCode.KO.IPCEventContext)_exContext).Process + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "IPC Connection String: " + ((SourceCode.KO.IPCEventContext)_exContext).ConnectionString + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.LineRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Start Activity Name: " + ((SourceCode.KO.LineRuleContext)_exContext).LineInstance.StartActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Finish Activity Name: " + ((SourceCode.KO.LineRuleContext)_exContext).LineInstance.FinishActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Line Name: " + ((SourceCode.KO.LineRuleContext)_exContext).LineInstance.Line.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Line Rule: " + ((SourceCode.KO.LineRuleContext)_exContext).LineRule.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.PrecedingRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.PrecedingRuleContext)_exContext).ActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Preceeding Rule: " + ((SourceCode.KO.PrecedingRuleContext)_exContext).PrecedingRule.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.ProcessEscalationAction:

                    case SourceCode.KO.ContextType.ProcessEscalationRule:

                    case SourceCode.KO.ContextType.ProcessFinishRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            break;

                        }

                    case SourceCode.KO.ContextType.ProcessStartRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Message: " + ((SourceCode.KO.ProcessStartRuleContext)_exContext).Message + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Start: " + ((SourceCode.KO.ProcessStartRuleContext)_exContext).Start.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            break;

                        }

                    case SourceCode.KO.ContextType.ServerEvent:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.ServerEventContext)_exContext).ActivityInstanceDestination.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Event Name: " + ((SourceCode.KO.ServerEventContext)_exContext).Event.Name + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.StartRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.StartRuleContext)_exContext).ActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            break;

                        }

                    case SourceCode.KO.ContextType.SucceedingRule:
                        {

                            _sMessage += "Exception thrown by: " + K2.ContextType.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            _exContext = K2.ContextObject;

                            _sMessage += "Activity Name: " + ((SourceCode.KO.SucceedingRuleContext)_exContext).ActivityInstance.Activity.Name + System.Environment.NewLine + System.Environment.NewLine;

                            _sMessage += "Succeeding Rule: " + ((SourceCode.KO.SucceedingRuleContext)_exContext).SucceedingRule.ToString() + System.Environment.NewLine + System.Environment.NewLine;

                            break;
                        }
                    default:
                        {
                            break;
                        }
                }

                _sMessage += System.Environment.NewLine + "Exception: " + System.Environment.NewLine;

                // Get the actual exception thrown
                ex = (Exception)K2.ExceptionObject;

                while (_bFlag == true)
                {
                    if (ex.InnerException != null)
                    {
                        ex = ex.InnerException;
                    }

                    else
                    {

                        _sMessage += "\tMessage: " + ex.Message + System.Environment.NewLine;

                        _sMessage += "Stack Trace:" + System.Environment.NewLine;

                        _sMessage += "\t" + "\t" + ex.StackTrace + System.Environment.NewLine + System.Environment.NewLine;

                        _bFlag = false;

                    }
                }

                _expMail.Body = _sMessage;

                System.Web.Mail.SmtpMail.Send(_expMail);

                // Rethrow the original exception to make sure the process goes into error state

                throw new Exception(ex.Message);

            }

            catch (Exception ex2)
            {
                throw new Exception(ex2.Message);
            }

            finally
            {
                ex = null;
                _exContext = null;
                _expMail = null;
            }

K2 Good Practice: SmartObject Naming and Organisation

Murphy — Fri, 20 Nov 2009 09:11:22 GMT

SmartObjects are just like any other item you will create in Visual Studio and you should be looking to make sure that you keep them well named and structured in your solutions. This will give you many benefits including:

1) Good naming will ensure that you do not confuse your SmartObjects with another with the same name in a shared platform. Consider if you have a SmartObject called Person in you project and you are deploying to a shared K2 server, it is very possible that someone with another project will also have a SmartObject called Person. As SmartObjects are uniquely identified by their name you now have a clash.

Consider using a namespace notation e.g ProjectName.Person

*Note when you do this if you are coding against the SmartObject API then the SmartObject name will be ProjectName_Person.

Sample Project Structure

2) Implement good organisation by placing your SmartObjects into folders in you project. This will

Ensure that if you have a large number of SmartObjects you project they are easy to find in Visual Studio.
When you deploy the SmartObject to K2 it will be placed in a category structure that matches your folder structure. This makes locating the SmartObject when using the tools like the SmartObject Tester easy.

SmartObjects as seen through the SmartObjectTester Tool

*Note that SmartObjects are placed in Categories directly related the project folder structure and not the project name, therefore if you want to change it then change the folder(s) of your project.

For further good practices make sure that you see the K2 blackpearl Best Practices knowledgebase article.

Simple batch scripts to restart your K2 Service, or to launch console mode

NeilM — Mon, 12 Oct 2009 14:02:00 GMT

Sometimes, it is necessary to recycle your K2 BlackPearl Service, for example when you want to update service type assemblies in the ServiceBroker folder.You could go into the services management console and stop and start the service, or you could create a small batch file that would do it for you.

This script will restart your K2 service - nte that thre is a built-in delay to wait for resources to be released, before starting the service again. Copy this text into a text file, and save it as a .bat file anywhere on your K2 server. When you run the batch file, your K2 service will be restarted.

REM SCRIPT STARTS HERE

@echo off
echo This batch will stop and start the K2 server service,
pause
echo Stopping the BlackPearl service (ignore errors stating that the service is stopped)
net stop "K2 blackpearl Server"
echo wait for 30 seconds to ensure resources are released
ping -n 30 localhost > nul
net start "K2 blackpearl Server"
echo if no errors were reported, you may close this window
pause

REM SCRIPT ENDS HERE

It is also quite common to run your K2 server in console mode during development. As you know, you should run your K2 server .exe process as the K2 Service account when in console mode, to ensure that the security context is the same as when the service is running in 'normal' mode.

Utilities such as BetterRunAs and PsExec allow yo to launch applications under differnet user credentials. The batch file example below assumes you have BetterRunAs copied on your local machine.

The script will stop the K2 server service, then launch the K2 server in console mode, and finally, when you have closed the console window, restart the K2 service

REM SCRIPT STARTS HERE

@echo off
rem ---------------------------
rem Usage: BetterRunAs.exe "application" "domain" "userid" "password"
rem ---------------------------
echo This batch will stop the K2 server service, and start K2 in console mode.
pause
echo Stopping the BlackPearl service (ignore errors stating that the service is stopped)
net stop "K2 blackpearl Server"
echo remember to restart the K2 service after you are finished
echo running it in console mode.
echo wait for 20 seconds to ensure resources are released
ping -n 20 localhost > nul
echo Start the blackpearl service as the service user in console mode
REM 64-bit systems: use the following line
REM "C:\Program Files (x86)\runas\BetterRunAs.exe" "C:\Program Files (x86)\K2 blackpearl\Host Server\Bin\K2HostServer.exe" [DOMAIN] [K2SERVICEACCOUNT] [K2SERVICEACCOUNTPASSWORD]
REM 32-bit Systems use the following line
"C:\Program Files\runas\BetterRunAs.exe" "C:\Program Files\K2 blackpearl\Host Server\Bin\K2HostServer.exe" [DOMAIN] [K2SERVICEACCOUNT] [K2SERVICEACCOUNTPASSWORD]
echo *If you wish, leave this window open, and once you close the K2 console window,
echo press enter to restart the K2 service automatically
echo (close the K2 console window first!)
pause
echo wait for 20 seconds to ensure resources are released
ping -n 20 localhost > nul
echo Starting the BlackPearl service
echo if any errors are reported, please check them
net start "K2 blackpearl Server"
echo if no errors were reported, you may close this window
pause

REM SCRIPT ENDS HERE

How to implement a ASP.NET Security Provider for K2 blackpearl

arno — Wed, 17 Jun 2009 11:18:00 GMT

The example below takes you through the steps to use the ASP.NET Authentication Provider with K2 blackpearl. A typical scenario where you might use the ASP.NET Security Provider is when your Web Application requires a user to login in order to view restricted content. This includes a K2 Worklist. This article shows how to take advantage of the out-of-box ASP.NET security components and how to extend it to enable successful integration with K2 blackpearl. The code ('ASPNetSecManager.zip') for this blog post can be downloaded from http://www.k2underground.com/groups/k2_blackpearl_custom_security_provider/media/p/31588.aspx. Please ensure you login the Underground before accessing the link. Please also note the accompanying readme.txt

Note: Full visibility of some images might be suppressed due to auto cropping. Simply right-click the image and paste the URL in a new browser window in order to view the complete image

Let’s get started:

1. Ensure K2 Host Server is stopped. (Start > Run >Services.msc, find ‘K2 blackpearl Server’, right click and select ‘All Tasks > Stop’)

2. Create a Folder where your solution will be created in. Let’s name this “ASPSec”

3. Create a containing folder for the Web Applications we will use. (We will create 2 Web Applications; one for the User Manager and one that will enable us to test the Workflow)

4. Create the 2 Web Application Folders

5. Extract the “ASPNetSecManager.zip” to a favourite location

6. Copy and Paste the contents of “www” to the folder indicated in the image

7. Copy and Paste the contents of “wwwSecManager” to the folder indicated in the image

8. Open SQL Management Studio and create a Blank database that will be extended with the ASP.NET Security Provider artifacts. Name it ‘ASPSec’

9. Create a new SQL Account that will be used to connect to the ASPSec database by the SecurityProvider

10. Ensure to grant the permissions below

(Tip: After creating the account, click on the ‘Connect’ icon in Management Studio and enter your newly created account details. This will ensure any SQL Policies (i.e. ‘Change account on next login’ issues can be resolved at this stage)

11. Let’s extend the database with the ASP.NET Membership artifacts. Open VS Command prompt

12. Run the aspnet_regsql command

13. Select the 1^st option and click Next

14. Select the Membership Store to extend and finish the wizard

15. Execute the following SQL Command against the ASPSec database

16. Let’s configure the User Manager and Workflow test Web Applications. I will do the Workflow test Web App first. Open IIS and configure a new virtual directory

17. Enter the details below. Note the name is the same as the value we inserted into the ‘aspnet_Applications’ table earlier

18. Enter the path to the source files

19. Set the Virtual Directory values

20. Set Anonymous access only.

21. Let’s configure the User Manager Web App now. Create a new virtual directory in IIS

22. Enter the virtual folder name below

23. Set the path

24. Set the Properties

25. Enable Integrated Authentication Only

26. Set the default content page to ‘Home.aspx’

27. Let’s create a blank Visual Studio Solution and add out Web Projects to it. We will also add an example K2 Security Provider project and a K2 Workflow Process to the solution

28. Let’s add the 2 Web Applications 1^st. Add an existing website

29. Browse and add the ‘Local IIS’ ‘ASPSecWeb’ Application

30. Repeat the steps for the ‘ASPSecMan’ Web Application

31. Now, let’s copy the Security Provider project to our solution folder

32. Do the same foe the Test K2 Process. Copy the code from the source folder

33. Create the destination folder and paste the contents into it

34. We are ready to add the Security Provider and the K2 Process to our solution. Add the Security Provider 1^st by adding an existing project and browsing to the solution folder where we pasted the files earlier

35. ...and the same for the K2 Process

36. Open ‘Home.aspx’ in the security manager and note the URL. If you create your own Web Application, ensure the settings are reflected in the querystring. This example manager is the one that ships with .NET and are used for demonstration purposes only.

37. Open the Web.Config of the ‘ASPSecWeb’. Note we configured the connectionstring node to include our SQL Login Details.

38. Let’s compile the Security provider now. Ensure you sign the Assembly. Save and compile the project

39. We need the Public Key token of the Compiled Security Assembly. Launch the VS Command Prompt and execute the ‘sn –T’ command to extract the value

40. Copy the ‘ArnIT.ASPNetSecurity.Library’ Assembly to the ‘securityproviders’ folder in the K2 Program files

41. With K2 Host Server still stopped, execute the following SQL Command against the HostServer database. Ensure your public key token is used in the script’s 3^rd parameter

42. Open the K2HostServer.config file and set the ‘useassemblyregistration’s value attribute to ‘true’

43. Start K2 Host Server. Host Server will register the new provider assembly and insert an entry into the ‘HostServer.SecurityProviders’ table automatically

44. Wait a few moments and execute the following SQL Statement against the HostServer database. This script is available from ‘regSQL.txt’ in the accompanied download zip file.

45. STOP K2 host server and open the K2HostServer.config file again.

46. Delete the encrypted connectionStrings section after you have learned the replacement values

47. Enter the replacement values

48. Enter the following Membership Specific section in the config file (you can copy this section from the ‘ASPSecWeb’s’ web.config file). Remember to save the config file only if you are happy that you have a backup and that the contents are correct!

49. So far so good. The registration is complete and you can start K2 HostServer again. If you start it in Console mode, you should see the provider registration information text during startup

Let’s test our work. We will firstly create a new user using our Membership Administration Web Application. We will then configure the K2 destination rule in our test K2 process to point to that user and we will test the lot using the ‘ASPSecWeb’ application by logging in using our user and starting a new process and auctioning a worklist item

50. Launch the Membership Administration Application

51. Click on ‘Create User’

52. Enter the user details and click ‘Create User’

53. Create an example role. Although we won’t use this, its good to demonstrate you can use the roles as well

54. Add our newly created user to the created role

55. Open the destination rule of the ‘CustomerResponse’ K2 process

56. Remove all the existing users

57. Add a new user. Note our users and groups are available in the Context Browser.

58. Add the user as indicated

59. Deploy the project

60. Assign ‘Start Permissions’ to our user

61. Finally, lets play with our implementation. Browse the ‘Default.aspx’ file in the ‘ASPSecWeb’ web application

62. Enter our user’s username and password. Note, this is the ASP.NET login control. You can use the ‘Password Reminder’, ‘Registration’ or any other controls in your application now.

63. Start a new Process by clicking on ‘Start Random Request’

64. Click on ‘View Worklist’ to inspect the result. You should have a new worklist item

65. Browse to ‘Instances’ in K2 Workspace and ensure the process has started. Note the ‘Originator’ value

Go ahead and finish the Worklist item.

A few final thoughts:

· You can use all the standard ASP.NET Membership controls (registration, password reminder ect.)

· The provider implementation is an example only, and you should re-visit the implementation to streamline the code

· Hosting of the Provider in K2 Host Server could be made more dynamic by taking advantage of the ‘AuthInit’ field in the SecurityLabels table in the HostServer database

· You can create your own Membership Administration UI and use the Microsoft Membership objects in your implementations

· A good idea is to create Membership Groups (i.e. Approvers) and use this is as source for your K2 Role that is used as part of your Destination designing in K2 Designer for Visual Studio. This ensures dynamic destinations

· You might want to implement a ‘trusted’ model between your Web Application and K2, which means you don’t pass the users password to the K2 Security provider for secondary authentication. If you do require secondary authentication, ensure all passwords are hashed and encrypted

Asynchronous server event processing between processes (Or, IPC's without the IPC-bit in the middle)

NeilM — Tue, 12 May 2009 21:36:00 GMT

So recently I had an interesting design challenge.I had a 'child' workflow which may or may not start before it's 'Parent' workflow, but at some stage, I needed the parent workflow to wait for the child workflow to complete.

Now I couldn't use a standard IPC event,since the parent workflow may not exist at the time the child workflow starts, and I also could not pass the K2 event serialnumber to the child process as one would normally do with an asynchronous server event, since the child workflow may be at some indeterminate point or not even exist, at the point when the parent workflow got to where it had to wait for the child to complete.

Hmmm...

If only there were some place to store the server event serial number, until the client event could pick up the serial number and finish the asynchronous server event task. Smartobjects to the rescue!

To make sense of all this, open up the attached K2 solution, and consider the two processes.

'AsynchronousProcess' is the parent process, and after starting, it will wait at the 'Wait here for a signal to continue' activity. In the activity, there is an event 'Add async server event entry'. Note the

K2.Synchronous = false;

line in this event - that is the bit that tells K2 not to compete the event, but to wait for an API call from outside to finish the event. This event creates the SmartObject record, and then waits for the child workflow to come around and tell it to complete.

The process 'ProcessThatCompletesAsyncProcess' then starts, and once you finish the client event 'Some client event', the 'child' workflow will go to the SmartObject, find any outstanding server items with a specific Unique Identifier, and then finish them off.

The nice thing about this approach is that there is no dependency on which process starts first, and nothing prevents the child workflow from finishing multiple parent workflows. As long as the child workflow can find the tasks it needs to finish (I use the 'Identifier' and 'Event Name' values in the SmartObject for this), you're all set.

Now, a couple of notes:

1) I made the sample as simple as possible, but you probably will want to delete or otherwise change the AsyncServerItems SmartObject record in the SmartBox database after the async server event has finished, to prevent errors when other child workflows find, and try to finish, a server event that no longer exists.

2) Yes,I know it is BAD practice to reference the workflow client .dll in a K2 process. In this case, it is OK though, since the child workflow will not attempt to interact with itself through the API, so the risk of deadlock is minimal. That said, a better option would probably be to call a asynchronous web service method to finish the server event, but again, I tried to keep the sample as simple as possible, And hey, it works :)

3) If it DOESN'T work, and you get the '26033 worklist item not open' error while running 807 update 370, you need to obtain and install the patch described in KB000384 (http://kb.k2workflow.com/articles/kb000384.aspx)

Feel free to have a play with the concept - you may find it useful in future...

Cheerio

Neil

K2 and Firewalled Domains

Murphy — Mon, 02 Mar 2009 16:37:00 GMT

Hi all I found myself involved in troubleshooting how to enable K2 to retrieve and authenticate users from a different domain sitting behind a firewall.

- The first step is easy and I am sure you all know, if you are running blackpearl 807 then you can open the workspace select ‘Management’ tab -> ‘User Managers’ -> ‘Domains’ and add your domain with a valid LDAP path.

- There are several pre requisites

o There must be trust between the domains

o The firewall must have enough ports open for the LDAP call to be made, the ports we found were required are:

§ RPC endpoint mapper: 135/tcp, 135/udp

§ RPC randomly allocated high TCP ports TCP 1024 - 65535*

§ * For more information about how to customize this port see http://support.microsoft.com/kb/224196/

§ 445/tcp, 445/udp

§ LDAP: 389/tcp/389/udp

§ Kerberos: 88/tcp, 88/udp

§ DNS: 53/tcp1, 53/udp

- Note: There is a good Microsoft KB articles which also covers ldap ports in more detail which is: http://support.microsoft.com/kb/832017/

Testing the Setup

If you run up the K2 workspace and browse for users in the alternate domain and you get a list back well that’s great you are off to a good start. If not then here are a few things you can do:

- Look in the blackpearl\hostserver\bin\ directory for a file called ADUM.txt, open this file to see if any errors have been logged and the detail.

- Get back to basics, do not use K2, run Microsoft’s ADSI edit and try and connect to the domain, this is a great idea as it moves the testing out of the K2 arena.

- Check your event logs, resolve any issues you may find

- Get the firewall logs (filter them to be relevant for traffic from the server you are using) and review what has been blocked and try and relate this back to whether it is a requirement for the LDAP lookup.

- Use a tool like PortQry to validate that the firewall ports you have asked to be opened can be communicated on (sometimes people think they have opened a port but have got it wrong, missed one ect..)

Once this is done, start a process which sends a client event to a user in this domain and test that they are able to open and action the worklist item.

Issues I came across:

- Kerberos port was not opened

- Kerberos packet fragmentation was occurring (see kb http://support.microsoft.com/kb/244474 )

- UDP port for LDAP was not opened, this needs to be opened as it looks like UDP is used to do the initial ping to find available DC’s ect..

Interesting point:

We were able to connect to the domain using ADSI edit if we specified specific user credentials as part of the connection string even without opening the Kerberos ports and LDAP UDP, however using integrated authentication failed. This is an important point to note as K2 will require integrated authentication and some tools may show they can retrieve users but this may only be due to the fact that they have entered specific credentials for the LDAP connection.

Well that is it from me, I hope that this post proves helpful to someone, also I am keen to receive comments on this blog, my knowledge in this area is limited and I am keen to be corrected, advised on any items in this post.

Load balanced Destinations with Smart objects (NO code)

Willem — Wed, 11 Feb 2009 20:16:00 GMT

Overview
A client of K2 asked me what load balancing capabilities we have with regards to destinations. The initial answer would typically be ‘None’ out of the box, but this is not entirely true.
We do have load balancing capabilities for destination groups Out-Of-The-Box: it is just not configured yet when you install K2.

In this document I will discuss load balancing capabilities with Smartobjects and destination rules without writing a line of code. I have created a pattern that utilizes stock standard ootb K2 Smartobjects with no need to write a service object.

In this article there are 3 Smartobjects included in this article and an InfoPath document to start and test some features.

Smartobjects

DestinationGroupBalanced.sodx
This is the Smartobject to be called from your process. The required input properties are the process name and the activity name for which you require (and have configured) load balanced groups.

Fields
Processname Identifies the process level grouping since each process has multiple activities, each of these may be required to be load balanced. Process name was also chosen as an identifier since the process name is available for selection from the object browser thus limiting any additional coding to pass in parameter detail
Activityname Per process each activity is unique. This is the second identifier required to determine which groups to assign for load balancing. As for process name, activity is available from the context browser
Groupname This is the actual destination rule name. Typically ‘mygroupname’ would be assigned to K2 destinations and therefore the name used in groupname should be a valid destination name.
Nextgroupid When using multiple groups and you want to split the load between those groups, each group must have the id of another group in the same ‘process activity’ pairing.
In the context of this Smartoject it is used internally and you should not concern yourself with this value during runtime.
Currentassignedgroupid This is the value returned from ‘NextGroupID.sodx’ and is used for filtering. Once again, do not reference this value during runtime.

LoadBalanceGroups.sodx
This Smartobject is used to return a single group name where the group name is the destination name

Fields:
Groupid Each group name in this object requires a unique id.
Groupname This is the actual destination name. this value will be assigned to the destination rule for an event.
Nextgroupid For multiple groups at an activity each group will indicate the next group to be assigned as the destination. Since all work is guaranteed to be distribiuted evenly with this pattern there is no need to worry about which group is first and which is last.
Care has to be taken though: The group names and the link (nextgroupid) shoud match the destinations required in the process activity pairing.

NextGroupID.sodx
This object tracks the last assigned group. Once the last assigned group id is returned by this object, an update is made to this object to indicate the next object assigned for a destination.

Fields:
Processname The process for which the next group name should be retrieved
Activityname The activity for the process where the next available group name is required
CurrentAssignedGroupID The group id of the group name to be used as the next destination

Data
LoadBalanceGroups and NextGroupID together determine the next group name to be returned.
Please study the following two data sets:

1- NextGroupID

2- LoadBalanceGroups

In the LoadBalanceGroups list Group One, Group Two, Group Three are all linked by the ‘Next Group ID’ field and the same for Group AAA and Group BBB.

In the NextGroupID list, the group with id ‘1’ will be assigned as the next destination, once Group One is returned, the ‘Next Group ID’ for Group One is used to update the ‘Last Assigned Group ID’ that will in turn become the next destination to be assigned.

When DestinationGroupBalanced is executed, Process name and activity name will now be passed and the relevant group name is returned to the interface.

Setup
1- To implement this pattern in your environment, deploy the smartobjects in this article to your environment.
2- You then have to integrate the smartobjects assigned to the InfoPath form. The current data sources will not work in your environment since my objects point to a new server. Refresh these references.
3- Add some dummy data via the InfoPath form to test the concept. Initial configured data is required.
4- Replace the dummy values above with actual destination names and test in a process.

If you cannot get the infopath form to work, use the SmartObject Service Tester.exe
([k2installdrive]\Program Files\K2 blackpearl\ServiceBroker) to populate LoadBalanceGroups and NextGroupID with data as above, then execute DestinationGroupBalanced

This pattern could cater for more complex scenarios but I think for a start it is not too bad.

Hope this helps

Willem Buchling

Collation error in the workspace after upgrading to 807

Keif — Wed, 19 Nov 2008 16:44:00 GMT

I've been working on a client site and one of the things I am working on is planning and testing for an upgrade to K2 [blackpearl] version 807 from version 803. The upgrade testing has gone smoothly so far except for an error we experienced in the workspace. The error is "cannot resolve the collation conflict between Latin1_General_CI_AS and SQL_Latin1_General_CP1_CI_AS in the like operation".

I did some investigation and this not a K2 [blackpearl] error. In effect this boiled down to the SQL server's collation being changed at some point in time after K2[blackpearl] was installed. There is a temporary table being used in generating a work-list in the workspace for 807 and these two facts combined to mean that the tempdb has a different collation to the columns in the K2 server DB and hence the error.

I decided the best resolution to this issue was to massage the collation on all the columns in the K2Server DB.

To do this I wrote a SQL script which I have attached to this blog post. All my initial testing with the client suggests this resolves the collation issue.

Hopefully anyone else running into this issue will benefit from this solution.

The attached script is supplied as is, I encourage you to back up any databases you apply this to and to test the script rigorously before applying this in production. The script is only attended to change the K2Server database, however the core could be adapter for other collation manipulate needs.

25th November, 2008 - Update:

I've update the attached script to a zip file containing scripts for all the K2 databases. I've excluded the SmartBox database as it's indexes are different for each customer's specific implementation, however the scripts supplied can form a model for which a custom SmartBox migration script can be created. I have also included a script for the Workspace database, however the index drop and create statements will need to be customised for your specific database as the index names are not the same for all implementations. I have also updated the scripts to include a statement to alter the default collation on the database itself and not just the columns within the database.

how to nicely end a workflow - programatically

chalama — Fri, 31 Oct 2008 16:21:00 GMT

You can find case where a user start a process that goes in sevral approval setps etc ... but he wants to withdraw (end) the process.

To do that you will need to design your process so it does have a "end" activity and we will use the Gotoactivity API function.

You will need the following reference:

using SourceCode.Hosting.Client.BaseAPI;
using SourceCode.Workflow.Management;

Here is a sample code that will do that :

try
        {
            // create a connection string
            SourceCode.Hosting.Client.BaseAPI.SCConnectionStringBuilder connBldr =
                    new SourceCode.Hosting.Client.BaseAPI.SCConnectionStringBuilder();

            // force the connection string to a specific credential of k2 server admin
            SCConnectionStringBuilder cb = new SCConnectionStringBuilder();
            cb.Host = "athena";
            cb.Port = 5555;
            cb.SecurityLabelName = "SH";
            cb.UserID = "orion";
            cb.WindowsDomain = "";
            cb.Password = "passe";
            cb.Integrated = false;
            cb.IsPrimaryLogin = true;

            // instantiate a connection with this connection string
            SourceCode.Workflow.Management.WorkflowManagementServer server = new    SourceCode.Workflow.Management.WorkflowManagementServer(cb.Host, cb.Port);
            server.CreateConnection();
            server.Connection.Open(cb.ToString());

            //
            foreach (SourceCode.Workflow.Management.ProcessSet ps in server.GetProcSets())
            {
                //if (ps.FullName == "K2WorkflowProject1\\Process1")
                {
                    Response.Write("Process : " + ps.FullName + "<br /> " );

                    foreach (SourceCode.Workflow.Management.ProcessInstance itemProc in server.GetProcessInstances(ps.ProcID))
                    {
                        //if (itemProc.Originator == "SH:sponsor")
                        {
                            Response.Write("proc ID = " + itemProc.ID.ToString() + " Folio =" + itemProc.Folio + " Originator = " + itemProc.Originator + "<br>");
                            if (itemProc.ID == 6){

                                //Connect to Workflow Server
                                SourceCode.Workflow.Client.Connection wfConnection = new SourceCode.Workflow.Client.Connection();
                                cb.Port = 5252;
                                wfConnection.Open("athena", cb.ToString());

                                //Get ProcessInstance
                                SourceCode.Workflow.Client.ProcessInstance proc = wfConnection.OpenProcessInstance(itemProc.ID);

                                proc.DataFields["EndByUser"].Value = "Sponsor";
                                proc.Update();
                                server.GotoActivity(itemProc.ID, "End");
                            }

                        }
                    }
                }
            }
            // close the connection
            server.Connection.Close();
        }
        catch(Exception ex)
        {
        }

Simple Dynamic Decision tree - example of how to implement it

chalama — Thu, 30 Oct 2008 10:02:00 GMT

Recently at client site I have been ask to create a dynamic decison tree.

A process that will be able to ask multiple departement to vote on a something knowing that the number of deprtement is unkonw before the start of the process and number of level of decision is also unkonw. The decison process needs to be run in paralell in all departement.

That decison process is looking like that :

Start

Departement 1/Approval Level 1 Departement 2/Approval Level 1 Departement 3/Approval Level 1 ----- Department N

Departement 1/Approval Level 2 Departement 3/Approval Level 2

Departement 3/Approval Level 3

Departement 3/Approval M

So we need to be able to dynamically creates N branches (one per department) and dynamically create M approval steps (one for each level of approval).

A solution for that will have 2 processes 1 to create parallel branches for each department and one that will handle the M approval steps.

Inside the K2 this will look like that:

The Parent process is called "NBranches" and the child process is called "Mapprovalsteps".

"NBranches" call the "Mapprovalsteps" using an IPC event and passing it the number of approval step to create:

That call must be in synch since we do want all the branches to complete their approval steps then collect all the final decision taken.

The parallel processing of the branches is done using a destination rule (plan per slot no destination) in which we specify a variable for the number of branch we want:

There are 2 others event in the only activity of the "NBranches" process see below:

The "Increment Branch counter" will increment locally a counter that will allow us to have the ID of the current Branch we are dealing with the code is at follow:

K2.ActivityInstanceDestination.DataFields["branchnum"].Value = Convert.ToInt32(K2.ProcessInstance.DataFields["branchescounter"].Value.ToString()) + 1;

K2.ProcessInstance.DataFields["branchescounter"].Value = K2.ActivityInstanceDestination.DataFields["branchnum"].Value;

Where the "branchnum" is a activity datafields and the "branchescounter" is a process datafields.

The "concatDecision" is saving the branch decision into the global decision variable, the code for that is the following:

K2.ProcessInstance.DataFields["decisions"].Value = K2.ProcessInstance.DataFields["decisions"].Value + "//" + K2.ActivityInstanceDestination.DataFields["retbranchdecision"].Value;

Where "decision" is the holder of all the decision taken by all the branches and "retbranchdecision" the decision taken by the current branches.

Here are the datafields for the "Nbranches" process :

Here is the data mapping when calling the "Mapprovalsteps" process:

The "Mapprovalsteps" process is looking like that:

The "Approval" is where the users takes decision , then if the if there is another level of approval the process goes to "Route approval to next approval step" and then generate another "Apprpoval" activity.

To decision to either end or loop back to "approval" is done using Line rules here are the rules:

Line rule that goes to the "End" activity.

Rule that goes the "Route approval to next approval step" activity.

The only thing that happen inside the "Route approval to next approval step" is the an operation that decrement a the number of approval step:

K2.ProcessInstance.DataFields["numberofDecisionStep"].Value = Convert.ToInt32(K2.ProcessInstance.DataFields["numberofDecisionStep"].Value.ToString()) - 1;

"numberofDecisionStep" is a process datafields.

In the "End" activity we populate the final decision so we can return the it to the parent process, hre is the code:

K2.ProcessInstance.DataFields["BranchDecision"].Value = "DecisionTakenByBranchNum=" + K2.ProcessInstance.DataFields["BranchNum"].Value.ToString ();

InfoPath – Designing with performance in mind

Murphy — Wed, 22 Oct 2008 14:04:00 GMT

Firstly apologies, there are no nice pictures in this blog just text but useful none the less I hope.

My colleagues and I have recently been involved in a project that requires the use of InfoPath and in particular Form Services (InfoPath running through MOSS form services to provide a browser based form). What’s new about that I hear you ask, well nothing really, however the functional requirements and data dependencies were very high and InfoPath struggled massively.

With this in mind I thought that I would share some of the key areas were we managed to make significant performance savings. Some are obvious and require little effort; others require more effort to implement but are worth the time.

1) Reduce the size of your images

Our graphics legend at the UK K2 offices managed to reduce the size of the images we were using by 60% without losing too much quality.

2) Limit the number of external calls to SmartObject methods (web services).

Each time you make a call to a web service this takes time to broker the transaction, apply any rules you have in the form and then re-render the form. It is possible that 3 web service calls will result in the page and your rules being run three times, with a large complex form this can be time consuming.

K2 provides a great mechanism help with this: Composite SmartObjects! If you are making several calls to load/save data (not lists of data) then why not create a Composite smart object that passes/returns all your data in one web service call. This will prevent unnecessary form renders.

3) Limit the use of Conditional formatting

InfoPath provides the ability to apply conditional formatting to page elements i.e. hide/make visible controls. Form services will not always be able to work out client side if it should apply the format and how. This can result in a post back to the server, again a costly operation if this is a large form.

4) Limit the number of cosmetic options you have in form

By this I mean items such as changing the heading text of a table based on a user selection, as again these can result in a post back.

5) Limit the amount of data held in the form

The more data your form holds the slower it will be to render, only return what you need. This is particularly true when showing data in repeating tables and sections as form services creates quite a lot of html to render these.

6) Be careful with Drop-Down list boxes

It is quite easy to bind a drop down list box to an external data source to get the list and forget the actual volume of data that may be returned. In particular if you select the option to ‘Show only entries with unique display names’. Form services uses a piece of script to manage this functionality, if the list is large this script takes a long time and is run each post back, this can cause massive delays. Filter your data server side to take this burden away from form services.

7) Upgrade the Browser

IE6 can produce a horrible flicker as the page renders, IE7 appears to be much better at rendering the page.

There are other items that you can consider but the above are a good starting point.

An additional question to ask: Is InfoPath correct for you? you should consider very carefully if InfoPath is right for your project, the initial perceived time savings soon disappear as the functional and performance requirements of the UI increase.

And one small thanks to Hoshy who pointed out I should not use ‘Designing Peformant InfoPath Forms’ as the title for my blog post as although the IT world use the term Performant a lot it is not actually a word J