Consider Spring MVC java web-application, which provides some REST API.
Let's say it has many methods, one of them is DELETE /api/foo/{id}, which obviously deletes foo entity from the DB with given id.
The problem is that due to big data in the DB, this operation is not immediate, so if client tries perform simultaneously multiply delete operations on same entity, say
DELETE /api/foo/123 x N times (by mistake in client software of course),
it causes some unpleasant side effects in the DB (you know, if you try delete same entity in several transactions, that's not generally nice).
My question is: what is the best practice in Spring MVC to prevent such situations?
I can certainly introduce synchronisation on Foo id in each such update method (PUT/DELETE). I will need to do it for all entities and all PUT/DELETE API methods though, which I really don't want to do. I suppose it should be some elegant and nice solution, how to perform such type of synchronisation on interceptor/servlet level, i.e. not on service of controller level.
I can also create specific interceptor and perform there waiting for duplicated requests (requests with same URL and parameters). But again, it doesn't sound as an elegant solution (until I will be ensured that it is not possible to configure in Spring MVC somehow in more beauty way).
That is a problem of concurrency that shall be handled by using the appropriate transaction and locking level. Unfortunately, there is no single size fits all way here and depending on your actual requirements, you could have to implement optimistic or pessimistic locking, as well as one of the possible transaction level (from no transaction at all to serializable transactions).
In general, handling such questions at the web level is a bad idea, because you will end in questions like what to do in on request wants to delete some data that another one is displaying at the same time? In SpringMVC, the common way is to use transactional methods in the service layer. Additionaly, you should declare an optimistic or pessimistic locking system in the persistence layer.
Optimistic layer normally give a higher throughput, at the cost of some transaction ending in exceptions. In that case, current best practices are now to report the problem to the user asking him/her to send his/her request again.
Related
Let's presume that we have an application "mail client" and a front-end for it.
If a user is typing a message or editing the subject or whatever, a rest call is made to update whatever the user was changing (e.g. the receivers) to keep the message in DRAFT. So a lot PUT's are happening to save the message. When closing the window, an update of every editable field happens at the same time. Hibernate can't handle this concurrency: Each of those calls retrieve the message, edit their own fields and try to save the message again, while the other call already changed it.
I know I can add a rest call to save all fields at the same time, but I was wondering if there is a cleaner solution, or a decent strategy to handle such cases (like for example only updating one field or some merge strategy if the object has already changed)
Thanks in advance!
The easiest solutions here would be to tweak the UI to either
Submit a single rest call during email submission that does all the tasks necessary.
Serialize the rest calls so they're chained rather than firing concurrently.
The concern I have here is that this will snowball at some point and become a bigger concurrency problem as more users are interacting with the application. Consider for the moment the potential number of concurrent rest calls your web infrastructure will have to support alone when you're faced with a 100, 500, 1000, or even 10000 or more concurrent users.
Does it really make sense to beef up the volume of servers to handle that load when the load itself is a product of a design flaw in the first place?
Hibernate is designed to handle locking through two mechanisms, optimistic and pessimistic.
Optimistic Way
Read the entity from the data store.
Cache a copy of the fields you're going to modify in temporary variables.
Modify the field or fields based on your PUT operation.
Attempt to merge the changes.
If save succeeds, you're done.
Should an OptimisticLockException occur, refresh the entity state from data store.
Compare cached values to the fields you must change.
If values differ, you can assert or throw an exception
If they don't differ, go back to 4.
The beautiful part of the optimistic approach is you avoid any form of deadlock happening, particularly if you're allowing multiple tables to be read and locked separately.
While you can use pessimistic lock options, optimistic locking is generally the best accepted way to handle concurrent operations as it has the least concurrency contention and performance impact.
We are building a product, so from performance point of view I need some help.
We are using complete Spring (MVC, JPA, Security etc..)
We have a requirement where say for a particular flow there can be 100 Business Rules getting executed at the same time. There can be n number of such flows and business rules.
These rules when executed actually fetch records from tables in database, these will contain few LAZILY INITIALIZED ENTITIES also.
I used Futures/Callables for multi threading purpose, but the problem is it fails to load LAZY variables. It gives Hibernate loading exception, probably some issue in TRANSACTIONAL not getting distributed in different threads.
Please let me know if there is any other way to approach?
if some Entity /Entity Collection is lazy fetched , and you are accessing it in another thread, you will face LazyInitialization exception, as lazy loaded entities can be accessed only within a transaction, and transaction wont span across, threads.
You can use DTO pattern, or if you are sharing an entity across threads, call its lazy initialized collections getter within the transaction so that they are fetched within transaction itself.
If you really need async processing then I suggest you use the Java EE specified way i.e. using JMS/MDB.
Anyways, since it seems like you want all the data to be loaded anyway for your processing. So Eager fetch all the required data and then submit the data for parallel processing. Or let each of the Tasks (Callables) fetch the data they require. Essentially, I am asking you to change your design so that the transactional boundaries doesn't cross multiple threads. Localize the boundaries within a single thread.
Currently we're creating a single EntityManager per request in our web app.
We're wanting to create a transaction at the beginning of the request and close it at the end to get automatic transaction support throughout the entire request.
Has anyone tried this? and will this introduce concurrency issues or is should we just wrap all of our update code in a separate transaction?
Note we're using JPA 1.0
For simple apps, this is enough (one transaction per request). But for more complex scenarios you will need to use transactions on more granular level.
In principle I always wrap my update transactions as needed.
Anyway your approach doesn't prevent you to open a new transaction in the middle of the request if needed. So it is OK by me.
And something about the possible concurrency issues...
If you can foresee issues, address them, otherwise I wouldn't worry too much about it. Keep it simple and don't get caught up in premature optimizations
I'm asking this question given my chosen development frameworks of JPA (Hibernate implementation of), Spring, and <insert MVC framework here - Struts 1, Struts 2, Spring MVC, Stripes...>.
I've been thinking a bit about relationships in my entity layer - for example I have an order entity that has many order lines. I've set up my app so that it eagerly loads the order lines for every order. Do you think this is a lazy way to get around the lazy initialization problems that I would come across if I was to set the fetch strategy to false?
The way I see it, I have the following alternatives when retrieving entities and their associations:
Use the Open Session In View pattern to create the session on each request and commit the transaction before returning the response.
Implement a DTO (Data Transfer Object) layer such that every DAO query I execute returns the correctly initialized DTO for my purposes. I don't really like this option much because in my experience I've found that it creates a lot of boilerplate copying code and becomes messy to maintain.
Don't map any associations in JPA so that every query I execute returns only the entities I'm interested in - this will probably require me to have DTOs anyway and will be a pain to maintain and I think defeats the purpose of having an ORM in the first place.
Eagerly fetch all (or most associations) - in the example above, always fetch all order lines when I retrieve an order.
So my question is, when and under what circumstances would you use which of these options? Do you always stick with one way of doing it?
I would ask a colleague but I think that if I even mentioned the term 'Open Session in View' I would be greeted with blank stares :( What I'm really looking for here is some advice from a senior or very experienced developer.
Thanks guys!
Open Session in View has some problems.
For example, if the transaction fails, you might know it too late at commit time, once you are nearly done rendering your page (possibly the response already commited, so you can't change the page !) ... If you had know that error before, you would have followed a different flow and ended up rendering a different page...
Other example, reading data on-demand might turn to many "N+1 select" problems, that kill your performance.
Many projects use the following path:
Maintain transactions at the business layer ; load at that point everything you are supposed to need.
Presentation layer runs the risk of LazyExceptions : each is considered a programming error, caught during tests, and corrected by loading more data in the business layer (you have the opportunity to do it efficiently, avoiding "N+1 select" problems).
To avoid creating extra classes for DTOs, you can load the data inside the entity objects themselves. This is the whole point of the POJO approach (uses by modern data-access layers, and even integration technologies like Spring).
I've successfully solved all my lazy initialization problems with Open Session In View -pattern (ie. the Spring implementation). The technologies I used were the exact same as you have.
Using this pattern allows me to fully map the entity relationships and not worry about fetching child entities in the dao. Mostly. In 90% of the cases the pattern solves the lazy initialization needs in the view. In some cases you'll have to "manually" initialize relationships. These cases were rare and always involved very very complex mappings in my case.
When using Open Entity Manager In View pattern it's important to define the entity relationships and especially propagation and transactional settings correctly. If these are not configured properly, there will be errors related to closed sessions when some entity is lazily initialized in the view and it fails due to the session having been closed already.
I definately would go with option 1. Option 2 might be needed sometimes, but I see absolutely no reason to use option 3. Option 4 is also a no no. Eagerly fetching everything kills the performance of any view that needs to list just a few properties of some parent entities (orders in tis case).
N+1 Selects
During development there will be N+1 selects as a result of initializing some relationships in the view. But this is not a reason to discard the pattern. Just fix these problems as they arise and before delivering the code to production. It's as easy to fix these problems with OEMIV pattern as it's with any other pattern: add the proper dao or service methods, fix the controller to call a different finder method, maybe add a view to the database etc.
I have successfully used the Open-Session-in-View pattern on a project. However, I recently read in "Spring In Practice" of an interesting potential problem with non-repeatable reads if you manage your transactions at a lower layer while keeping the Hibernate session open in the view layer.
We managed most of our transactions in the service layer, but kept the hibernate session open in the view layer. This meant that lazy reads in the view were resulting in separate read transactions.
We managed our transactions in our service layer to minimize transaction duration. For instance, some of our service calls resulted in both a database transaction and a web service call to an external service. We did not want our transaction to be open while waiting for a web service call to respond.
As our system never went into production, I am not sure if there were any real problems with it, but I suspect that there was the potential for the view to attempt to lazily load an object that has been deleted by someone else.
There are some benefits of DTO approach though. You have to think beforehand what information you need. In some cases this will prevent you from generating n+1 select statements. It helps also to see where to use eager fetching and/or optimized views.
I'll also throw my weight behind the Open-Session-in-View pattern, having been in the exact same boat before.
I work with Stripes without spring, and have created a manual filter before that tends to work well. Coding transaction logic on the backend turns messy really quick as you've mentioned. Eagerly fetching everything becomes TERRIBLE as you map more and more objects to each other.
One thing I want to add that you may not have come across is Stripersist and Stripernate - Stripersist being the more JPA flavor - auto-hydration filters that take a lot of the work off your shoulders.
With Stripersist you can say things like /appContextRoot/actions/view/3 and it will auto-hydrate the JPA Entity on the ActionBean with id of 3 before the event is executed.
Stripersist is in the stripes-stuff package on sourceforge. I now use this for all new projects, as it's clean and easily supports multiple datasources if necessary.
Does the Order and Order Lines compose a high volume of data? Do they take part in online processes where real-time response is required? If so, you might consider not using eager fetching - it does make a huge diference in performance. If the amount of data is small, there is no problem in eager fetching.
About using DTOs, it might be a viable implementation.
If your business layer is used internally by your own application (i.e a small web app and its business logic) it'd probably be best to use your own entities in your view with open session in view pattern since it's simpler.
If your entities are used by many applications (i.e a backend application providing a service in your corporation) it'd be interesting to use DTOs since you would not expose your model to your clients. Exposing it could mean you would have a harder time refactoring your model since it could mean breaking contracts with your clients. A DTO would make that easier since you have another layer of
abstraction. This can be a bit strange since EJB3 would theorically eliminate the need of DTOs.
I'm thinking about using the Open Session In View (OSIV) filter or interceptor that comes with Spring, as it seems like a convenient way for me as a developer. If that's what you recommend, do you recommend using a filter or an interceptor and why?
I'm also wondering how it will mix with HibernateTemplate and if I will lose the ability to mark methods as #Transactional(readOnly = true) etc and thus lose the ability to get some more fine grained transaction control?
Is there some kind of best practice for how to integrate this kind of solution with a three tier architecture using Hibernate and Spring (as I suppose my decision to use Wicket for presentation shouldn't matter much)?
If I use OSIV I will at least never run into lazy loading exceptions, on the other hand my transaction will live longer before being able to commit by being uncommitted in the view as well.
It's really a matter of personal taste.
Personally, I like to have transaction boundaries at the service layer. If you start thinking SOA, every call to a service should be independent. If your view layer has to call 2 different services (we could argue that this is already a code smell) then those 2 services should behave independently of each other, could have different transaction configurations, etc... Having no transactions open outside of the services also helps make sure that no modification occurs outside of a service.
OTOH you will have to think a bit more about what you do in your services (lazy loading, grouping functionalities in the same service method if they need a common transactionality, etc ...).
One pattern that can help reduce lazy-loading error is to use Value Object outside of the service layer. The services should always load all the data needed and copy it to VOs. You lose the direct mapping between your persistent objects and your view layer (meaning you have to write more code), but you might find that you gain in clarity ...
Edit: The decision will be based on trade offs, so I still think it is at least partly a matter of personal taste. Transaction at the service layer feels cleaner to me (more SOA-like, the logic is clearly restrained to the service layer, different calls are clearly separated, ...). The problem with that approach is LazyLoadingExceptions, which can be resolved by using VO. If VO are just a copy of your persistent objects, then yes, it is clearly a break of the DRY principle. If you use VO like you would use a database View, then VO are a simplification of you persistent objects. It will still be more code to write, but it will make your design clearer. It becomes especially useful if you need to plug some authorization scheme : if certain fields are visible only to certain roles, you can put the authorization at the service level and never return data that should not be viewed.
If I use OSIV I will at least never
run into lazy loading exceptions
that is not true, in fact its extremely easy to run into the infamous LazyInitializationException, just load an object, and try to read an attribute of it, after the view, depending on your configuration you WILL get the LIE