I'm trying to work out how to resume a persisted workflow that I'm self-hosting with a WorkflowServiceHost. Currently my host wires up persistence and idling behaviour like so:
// Persistence
var connStr = #"";
var behavior = new SqlWorkflowInstanceStoreBehavior(connStr);
behavior.InstanceCompletionAction = InstanceCompletionAction.DeleteNothing;
behavior.InstanceLockedExceptionAction = InstanceLockedExceptionAction.AggressiveRetry;
behavior.InstanceEncodingOption = InstanceEncodingOption.None;
host.Description.Behaviors.Add(behavior);
// Idle behaviour
var idleBehavior = new WorkflowIdleBehavior();
idleBehavior.TimeToPersist = TimeSpan.FromMinutes(2);
idleBehavior.TimeToUnload = TimeSpan.FromMinutes(2);
host.Description.Behaviors.Add(behavior);
I am storing the workflow instance GUID against a ProductID in my database inside a custom activity contained in the workflow, so I am able to easily trace a specific workflow instance to a product ID. I want to be able to somehow pass this Product ID to my ServiceHost and have it resume the correct persisted workflow for me.
Can anyone point me in the right direction on how to do this?
Many thanks in advance
Ian
You're going to want to read about Content Correlation. You'll correlate your ProductID with your Receive activities. A bookmark will be created when you reach a receive, and your workflow will be persisted. You call the receive and pass your ProductID in and because of the correlation, the WF runtime knows which instance to resume.
Related
Is it possible to Add a message to an Azure queue then, in the same flow, update or delete that message?
The idea would be to use the queue to ensure that some work gets done - there's a worker role monitoring that queue. But, the Web role which added the message may be able to make some progress toward (and sometimes even to complete) the transaction.
The worker would already be designed to handle double-delivery and reprocessing partially handled messages (from previous, failed worker attempts) - so there isn't a technical problem here, just time inefficiency and some superfluous storage transactions.
So far it seems like adding the message allows for a delivery delay, giving the web role some time, but doesn't give back a pop-receipt which it seems like we'd need to update/delete the message. Am I missing something?
It seems this feature was added as part of the "2016-05-31” REST API
we now make pop receipt value available in the Put Message (aka Add Message) response which allows users to update/delete a message without the need to retrieve the message first.
I suggest you follow these steps as it worked for me
How to: Create a queue
A CloudQueueClient object lets you get reference objects for queues. The following code creates a CloudQueueClient object. All code in this guide uses a storage connection string stored in the Azure application's service configuration. There are also other ways to create a CloudStorageAccount object. See CloudStorageAccount documentation for details.
// Retrieve storage account from connection string
CloudStorageAccount storageAccount = CloudStorageAccount.Parse(
CloudConfigurationManager.GetSetting("StorageConnectionString"));
// Create the queue client
CloudQueueClient queueClient = storageAccount.CreateCloudQueueClient();
Use the queueClient object to get a reference to the queue you want to use. You can create the queue if it doesn't exist.
// Retrieve a reference to a queue
CloudQueue queue = queueClient.GetQueueReference("myqueue");
// Create the queue if it doesn't already exist
queue.CreateIfNotExists();
How to: Insert a message into a queue
To insert a message into an existing queue, first create a new CloudQueueMessage. Next, call the AddMessage method. A CloudQueueMessage can be created from either a string (in UTF-8 format) or a byte array. Here is code which creates a queue (if it doesn't exist) and inserts the message 'Hello, World':
// Retrieve storage account from connection string.
CloudStorageAccount storageAccount = CloudStorageAccount.Parse(
CloudConfigurationManager.GetSetting("StorageConnectionString"));
// Create the queue client.
CloudQueueClient queueClient = storageAccount.CreateCloudQueueClient();
// Retrieve a reference to a queue.
CloudQueue queue = queueClient.GetQueueReference("myqueue");
// Create the queue if it doesn't already exist.
queue.CreateIfNotExists();
// Create a message and add it to the queue.
CloudQueueMessage message = new CloudQueueMessage("Hello, World");
queue.AddMessage(message);
For more details, refer this link.
http://azure.microsoft.com/en-us/documentation/articles/storage-dotnet-how-to-use-queues/
Girish Prajwal
We are developing using VS2010 and MVC4, deploying our web app on an IIS 7.5 on Windows7.
Our project has a long running process for which we want to display status and progress.
In order to accomplish this we have a small serializable class with properties that describe the current status. The long operation pseudo code goes like this:
int curentPercentComplete = 0;
EngineStatus status = new EngineStatus();
while (!done) {
status.PercentComplete = curentPercentComplete;
Session['status'] = status;
// do lengthy operation
curentPercentComplete = compute();
done = isJobFinished();
}
We also have an other controller action that tries to retrieve the current status from the session
which then encodes to json and returns it to the browser via an Ajax request.
Our problem is that we always seem to get the last saved data from the previous request, in other words the session object does not seem to update the Session['status'] field during the execution of the while block.
We have tried the session state mode both InProc and StateServer with exactly the same behavior.
Thanks in advance.
It turns out that the MVC framework performs a single update of the session data at the end of the request which means that only the last value is saved.
Since the "lengthy" operation is performed in a single request-response cycle, the idea of storing intermediate status information in the session is plain wrong.
I have a Windows service using NServiceBus to handle incoming messages.
While processing a message, I would like to check to see if there are any other remaining messages on the queue to process.
What is the best way to approach this?
For this specific scenario I'd say that a saga could be appropriate where it is created by the first message received, opens a timeout (for let's say one minute), collects all messages during that period of time, then Bus.SendLocal's a message containing all rows, for which another handler creates the spreadsheet and uploads.
Since, NServiceBus is using MSMQ, you can use the methods from System.Messaging.
Included is a modified method, I'm currently working on, to do a kind of batch processing.
using System.Messaging;
public int PeekAtQueue()
{
const string QUEUE_NAME = "private$\\you_precious_queuname";
if (!MessageQueue.Exists(".\\" + QUEUE_NAME))
return 0;
var messageQueues = MessageQueue.GetPrivateQueuesByMachine(Environment.MachineName);
var queue = messageQueues.Single(x => x.QueueName == QUEUE_NAME);
return queue.GetAllMessages().Count();
}
Modified here itself in the editor. Hope it still compiles :)
Found a similar discussion here, by the way:
http://jopinblog.wordpress.com/2008/03/12/counting-messages-in-an-msmq-messagequeue-from-c/
I'm logging using log4net, and I want to log a id that is unique for each serice method call. I dont need it unique across service calls, just within a method call. Is there any built in id i can use in wcf? I don't want to manually create a guid or something at the start of the method call.
e.g.
wcfMethod(int x)
{
log("xxx");
somework
log("yyy");
}
private log(string message)
{
var frame = new StackFrame(1);
var method = frame.GetMethod();
var type = method.DeclaringType;
var name = method.Name;
var log = LogManager.GetLogger(type);
// LOG ID HERE
ThreadContext.Properties["MessageId"] = OperationContext.Current.IncomingMessageHeaders.MessageId; // SOMETHING HERE
}
I've tried OperationContext.Current.IncomingMessageHeaders.MessageId but thats always null.
I've read about wcf instance correlation but i don't need something that complicated (e.g. unique across different method calls).
Please if anyone can help that would be much apprieciated. Thanks in advance.
Plain SOAP or REST has no such identification included in messages. You must use some additional feature or transport protocol (for example MSMQ) supporting identifications of messages. In case of MessageId you have to use SOAP service with WS-Addressing and this information must be passed from client.
I am using the WCF 4 routing service, and need to configure the service programmatically (as opposed to via config). The examples I have seen of doing so, which are rare, create a MessageFilterTable as follows:
var filterTable=new MessageFilterTable<IEnumerable<ServiceEndpoint>>();
But, the generic parameter to that method is supposed to be TFilterData (the type of data you are filtering on)? I have my own custom filter that accepts a string -- can I still create the filter table this way?
If this will work...will the routing infrastructure create client endpoints out of the list I pass in?
I have created a WCF 4 routing service and configured it programmatically. My code is a bit more spaced out than it needs to be (maintainability for others being a concern, hence the comments), but it definitely works. This has two filters: one filters some specific Actions to a given endpoint, and the second sends the remaining actions to a generic endpoint.
// Create the message filter table used for routing messages
MessageFilterTable<IEnumerable<ServiceEndpoint>> filterTable = new MessageFilterTable<IEnumerable<ServiceEndpoint>>();
// If we're processing a subscribe or unsubscribe, send to the subscription endpoint
filterTable.Add(
new ActionMessageFilter(
"http://etcetcetc/ISubscription/Subscribe",
"http://etcetcetc/ISubscription/KeepAlive",
"http://etcetcetc/ISubscription/Unsubscribe"),
new List<ServiceEndpoint>()
{
new ServiceEndpoint(
new ContractDescription("ISubscription", "http://etcetcetc/"),
binding,
new EndpointAddress(String.Format("{0}{1}{2}", TCPPrefix, HostName, SubscriptionSuffix)))
},
HighRoutingPriority);
// Otherwise, send all other packets to the routing endpoint
MatchAllMessageFilter filter = new MatchAllMessageFilter();
filterTable.Add(
filter,
new List<ServiceEndpoint>()
{
new ServiceEndpoint(
new ContractDescription("IRouter", "http://etcetcetc/"),
binding,
new EndpointAddress(String.Format("{0}{1}{2}", TCPPrefix, HostName, RouterSuffix)))
},
LowRoutingPriority);
// Then attach the filter table as part of a RoutingBehaviour to the host
_routingHost.Description.Behaviors.Add(
new RoutingBehavior(new RoutingConfiguration(filterTable, false)));
You can find a good example on MSDN here: How To: Dynamic Update Routing Table
Note how they dont directly create an instance of the MessageFilterTable, but instead use the 'FilterTable' property provided by a new RoutingConfiguration instance.
If you have written a custom filter, then you will add it like this:
rc.FilterTable.Add(new CustomMessageFilter("customStringParameter"), new List<ServiceEndpoint> { physicalServiceEndpoint });
The CustomMessageFilter will be your filter, and the "customStringParameter" is the string that (I believe) you are talking about.
When the Router receives a connection request, it will attempt to map it via this table entry, if this is successful, then you are right, the router will create a client endpoint to talk to the ServiceEndpoint that you provided.