I believe we could benefit from caching our SagaStore.
So, to configure the CachingSagaStore.
The cache is written through, which covers moving from node-1 to node-2.
What happens though when the StreamingEventProcessor moves back to node-1 and the state was changed on node-2?
All associations are set at saga's start.
Application context:
Axon Framework version 4
Two application nodes
Four segments for our sagas
Found it, reading the javadoc will help..
"The tracking token of the event last handled by the Saga is also passed as a parameter."
...
Hmm that does not solve it.. a) the token is not passed, b) it would not solve the issue :- )
caching is hard!
Related
I am using an Axon Event Tracking processor. Sometimes events take longer that 10 seconds to process.
This seems to cause the message to be processed again and this appears in the log "Releasing claim of token X/0 failed. It was owned by another node."
If I up the number of segments it does not log this BUT the event is still processed twice so I think this might be misleading. (I think I was mistaken about this)
I have tried adjusting the fetchDelay, cleanupDelay and tokenClaimInterval. None of which has fixed this. Is there a property or something that I am missing?
Edit
The scenario taking longer than 10 seconds is making a HTTP request to an external service.
I'm using axon 4.1.2 with all default configuration when using with Spring auto configuration. I cannot see the Releasing claim on token and preparing for retry in [timeout]s log.
I was having this issue with a single segment and 2 instances of the application. I realised I hadn't increased the number of segments like I thought I had.
After further investigation I have discovered that adding an additional segment seems to have stopped this. Even if I have for example 2 segments and 6 applications it still doesn't reappear, however I'm not sure how this is different to my original scenario of 1 segment and 2 application?
I didn't realise it would be possible for multiple threads to grab the same tracking token and process the same event. It sounds like the best action would be to put an idem-potency check before the HTTP call?
The Releasing claim of token [event-processor-name]/[segment-id] failed. It was owned by another node. message can only occur in three scenarios:
You are performing a merge operation of two segments which fails because the given thread doesn't own both segments.
The main event processing loop of the TrackingEventProcessor is stopped, but releasing the token claim fails because the token is already claimed by another thread.
The main event processing loop has caught an Exception, making it retry with a exponential back-off, and it tries to release the claim (which might fail with the given message).
I am guessing it's not options 1 and 2, so that would leave us with option 3. This should also mean you are seeing other WARN level messages, like:
Releasing claim on token and preparing for retry in [timeout]s
Would you be able to share whether that's the case? That way we can pinpoint a little better what the exact problem is you are encountering.
By the way, very likely you have several processes (event handling threads of the TrackingEventProcessor) stealing the TrackingToken from one another. As they're stealing an un-updated token, both (or more) will handled the same event. Hence why you see the event handler being invoked twice.
Obviously undesirable behavior and something we should resolve for you. I would like to ask you to provide answers to my comments under the question, as right now I have to little to go on. Let us figure this out #Dan!
Update
Thanks for updating your question #dan, that's very helpful.
From what you've shared, I am fairly confident that both instances are stealing the token from one another. This does depend though on whether both are using the same database for the token_entry table (although I am assuming they are).
If they are using the same table, then they should "nicely" share their work, unless one of them takes to long. If it takes to long, the token will be claimed by another process. This other process in this case is the thread of the TEP of your other application instance. The "claim timeout" is defaulted to 10 seconds, which also corresponds with the long running event handling process.
This claimTimeout is adjustable though, by invoking the Builder of the JpaTokenStore/JdbcTokenStore (depending on which you are using / auto wiring) and calling the JpaTokenStore.Builder#claimTimeout(TemporalAmount) method. And, I think this would be required on your end, giving the fact you have a long running operation.
There are of course different ways of tackling this. Like, making sure the TEP is only ran on a single instance (not really fault tolerant though), or offloading this long running operation to a schedule task which is triggered by the event.
But, I think we've found the issue at least, so I'd suggest to tweak the claimTimeout and see if the problem persists.
Let us know if this resolves the problem on your end #dan!
Trying to work out why some of my application servers have creeped up over 1s response times using newrelic. We're using WebApi 2.0 and MVC5.
As you can see below the bulk of the time is spent under 'WebTransaction'. The throughput figures aren't particularly high - what could be causing this, and what are the steps I can take to reduce it down?
Thanks
EDIT I added transactional tracing to this function to get some further analysis - see below:
Over 1 second waiting in System.Web.HttpApplication.BeginRequest().
Any insight into this would be appreciated.
Ok - I have now solved the issue.
Cause
One of my logging handlers which syncs it's data to cloud storage was initializing every time it was instantiated, which also involved a call to Azure table storage. As it was passed into the controller in question, every call to the API resulted in this instantiate.
It was a blocking call, so it added ~1s to every call. Once i configured this initialization to be server life-cycle wide,
Observations
As the blocking call was made at the time of the Controller being build (due to Unity resolving the dependancies at this point) New Relic reports this as
System.Web.HttpApplication.BeginRequest()
Although I would love to see this a little granular, as we can see from the transactional trace above it was in fact the 7 calls to table storage (still not quite sure why it was 7) that led me down this path.
Nice tool - my new relic subscription is starting to pay for itself.
It appears that the bulk of time is being spent in Account.NewSession. But it is difficult to say without drilling down into your data. If you need some more insight into a block of code, you may want to consider adding Custom Instrumentation
If you would like us to investigate this in more depth, please reach out to us at support.newrelic.com where we will have you account information on hand.
I have two workflow services (state machines) that should cooperate and exchange information to accomplish the desired behavior.
The problem I have (but I also had it with only one state machine) is that sometimes I try to send an operation which is not allowed by the current state.
There are two problems: 1) I have to wait the operation timeout to know that the operation was not allowed 2) I'm "masking" real timeouts due to other problems
By now, I found two possible solutions: 1) I can change signatures to return true (allowed) and false (not allowed) and add all operations to all states, (not allowed operations would trigger a self-transition) 2) I always add all transitions to all states (not allowed would trigger a self-transition) but for transitions not allowed I will send an exception
I would like to know which is the best option (and, of course, I'd appreciate other possible solutions too).
I would also like to know how I could reply to a request with an exception (maybe throwing it within a try/catch?).
Thanks
Another option here is to use the information in the workflow persistence store. One of the columns contains the active bookmarks and in the case of a Receive activity this is the SOAP operation. You could have a separate service that exposes that information for a given workflow instance.
You still need to cater of the fact that you might send messages to a workflow that is in a different state because the workflow persistence store isn't updated right away (unless you make it do so) and because multiple people might send messages to the same workflow instance. Still this basic technique works really well and I have used this to enable/disable buttons on the UI based on the state of a workflow.
I know that similar questions have been asked all over the place, but I'm having trouble finding one that relates directly to what I'm after.
I have a website where a user uploads a data file, then that file is transformed and imported into SQL. The file could be up to 50mb in size, and some times this process can take 30 minutes or sometimes even longer.
I realise I need to palm off the actual work to another process, and poll that process on the web page. I'm wondering what the best approach would be though? Being a web developer by trade, I'm finding all this new Windows Service stuff a bit confusing, and I just wanted somewhere to start.
So:
Can I do / should I being doing this with a windows service? if so, how?
Should I use WCF? If this runs under IIS, will I have problems with aspnet_wp.exe recycling and timing out my process?
clarifications
The data is imported into sql, there's no file distribution taking place.
If there is a failure, it absolutely MUST be reported to the user. The web page will poll every, lets say, 5 seconds, from the time the async task begins, to get the 'status' of the import. Once it's finished another response will tell the page to stop polling for status updates.
queries on final decision
ok, so as I thought, it seems that a windows service is the best idea. So as to HOW to get it to work, it seems the 'put the file there and wait for the service to pick it up' idea is the generally accepted way, is there a way I can start a process run by the service, without it having to constantly be checking a database table / folder? As I said earlier, I don't have any experience with Windows Services - I wondered if I put a public method in the service, can I call it somehow?
well ...
var thread = new Thread(() => {
// your action
});
thread.Start();
but you will have problems with that:
what if the import to sql fails? should there be any response to the client
if it fails, how do you ensure the file on a later request
what if the applications shuts down ... this newly created and started thread will be killed either
...
it's not always a good idea to store everything in sql (especially files...). if you want to make the file available to several servers why not distribute them via ftp ...?
i believe that your whole concept is a bit messed up (sry assuming this), and it might be helpful if you elaborate and give us more information about your intentions!
edit:
Can I do / should I being doing this
with a windows service? if so, how?
you can :) i advise you to create a simple console-program and convert this with srvany and sc. you can get a rough overview howto here (note: insert blanks after =... that's a silly pitfall)
the term should is relative, because you did not answer the most important question
what if a record is persisted to the database, telling a consumer that file test.img should be persisted, but your service hasn't captured it or did not transform it yet?
so ... next on
Should I use WCF? If this runs under IIS, will I have problems with aspnet_wp.exe recycling and timing out my process?
you probably could create a WCF-service which recieves some binary-data and then stores this to a database. this request could be async. yes. but what for?
once again:
please give us more insight to your workflow: what are you exactly trying to achieve? which "environmental-conditions" to you have (eg. app A polls db and expects file-records which are referenced in table x to be persisted) ...
edit:
so you want to import a .csv-file. well that changes everything :)
but i won't advise you to use a wcf-service (there could be a usage: eg. a wcf-service which has a method to insert a single row, then your iteration through the file would be implemented in another app... not that good, though).
i would suggest following:
at first do everything in your webapp (as you've already done), but rather use some sort of bulk-insert and do your transformation/logic on the database.
if you have some sort of bottle-neck then, i would suggest you something like a minor job-service, eg:
webapp will upload the file and insert a row to a job-table. the job-service is continiously polling the table/or gets informed via wcf by the webapp (hey, hey, finally some sort of usage for WCF in your scenario... :) ) and then does the import-job, writing a finish-note to a table/or set the state of the job to finished ...
but this is a bit overkill :)
Please see if my below comments helps you to resolve your issue:
•Can I do / should I being doing this with a windows service? if so, how?
Yes you can do this with a windows service. And I think that is the way you should be doing it. You can implement your own service to process your request or you can use the open source code Job Proccessor
Basically the idea is..
You submit a request for processing
the csv file in database table with
some status as not started.
Then your windows service picks up
the request from database table which
are not started and update them as in
progress status.
Once the processing is complete
succesfully /unsuccesfuly your
service updated the database table
with status as Completed / Failed.
And your asp.net page can poll to
database table for the current status
every 5 sec or so.
•Should I use WCF? If this runs under IIS, will I have problems with aspnet_wp.exe recycling and timing out my process?
you should not be using WCF for this purpose.
I am thinking on the following approach but not sure if its the best way out:
step1 (server side): A TaskMangaer class creates a new thread and start a task.
step2 (server side): Store taskManager object reference into the cache for future reference.
step3 (client side): Use periodic Ajax call to check the status of the task.
Basically the intention is to have a framework to run a background task (5mins approx) and provide regular feedback on the web UI for the percentage of task completed.
Is there a neat way around this or any existing asp.net API that will be helpful ?
Edit 1#: I want to run the task in-proc with the app.
Edit 2#: Looks like badge implementation on stack overflow is also using the cache to track background task. https://blog.stackoverflow.com/2008/07/easy-background-tasks-in-aspnet/
I think the problem with storing the result in the cache is that ASP.NET might scavenge that cache entry for other purposes (ie if its short on memory, if its grumpy, etc). Something that is served from the cache should be something you can recreate on demand if its not found in the cache, the ASP.NET runtime is free to dump cache entries whenever it feels like it.
The usage of the cache in the badge discussion seems fundamentally different, in that case the task was shortlived. The cache was just being used as a hacky timer to fire off the task periodically.
Can you confirm this is a task that is going to take 5 minutes, and require its own thread that whole time? This is a performance concern in itself, you will only be able to support a limited number of such requests if each requires its own thread for so long. Only if thats acceptable would I let the task camp a thread for so long.
If its ok for these tasks to camp a thread, then I'd just go ahead and store the result in a dictionary global to the process. The key of the dictionary would correlate to the client request / AJAX callback series. The key should incorporate the user ID as well if security is at all important.
If you need to scale up to many users, then I think you need to break the task down into asynchronous steps, and in that case I'd probably use a DB table to store the results (again keyed per request / user).
Microsoft Message Queuing was built for scenarios like the one you try to solve:
http://www.microsoft.com/windowsserver2003/technologies/msmq/default.mspx
Windows Communicatio Foundation also has message queuing support.
Hope this helps.
Thomas
One approach for doing this is to use application state. When you spawn a worker thread, pass it a request ID that you generate, and return this to the client. The client will then pass that request ID back to the server in its AJAX calls. The server will then fetch the status using the request ID from application state. (The worker thread would be updating the application state based on its status).
I saw an approach to a similar problem somewhere. The solution was something like:
Start the background task on server.Return immediately with a url to the result.
Until the result is posted, this url will return 404.
The client checks periodically for this url.
The client reads the results when
they are finally posted.
The url will be something like http://mysite/myresults/cffc6c30-d1c2-11dd-ad8b-0800200c9a66.
The best document format is probably JSON.
If feedback on progress is important, modify the document to also contain status (inprogress/finish) and progress (42 %).