How to load @Cache on startup in spring?
Solution 1:
Just use the cache as before, add a scheduler to update cache, code snippet is below.
@Service
public class CacheScheduler {
@Autowired
BookDao bookDao;
@Autowired
CacheManager cacheManager;
@PostConstruct
public void init() {
update();
scheduleUpdateAsync();
}
public void update() {
for (Book book : bookDao.findAll()) {
cacheManager.getCache("books").put(book.getIsbn(), book);
}
}
}
Make sure your KeyGenerator
will return the object for one parameter (as default). Or else, expose the putToCache
method in BookService
to avoid using cacheManager directly.
@CachePut(value = "books", key = "#book.isbn")
public Book putToCache(Book book) {
return book;
}
Solution 2:
I have encountered the following problem when using @PostConstruct: - even though the method I wanted to be cached was called, after calling it from swagger, it still didn't use the cached value. Only after called it once more.
That was because @PostConstruct is too early for caching something. (At least I think that was the issue)
Now I'm using it more late in the startup process and it works without problems:
@Component
public class CacheInit implements ApplicationListener<ApplicationReadyEvent> {
@Override
public void onApplicationEvent(ApplicationReadyEvent event) {
//call service method
}
}
Solution 3:
An option would be to use the CommandLineRunner
for populating the cache on startup.
From official CommandLineRunner documentation, it is an:
Interface used to indicate that a bean should run when it is contained within a SpringApplication.
Hence, we just need to retrieve the list of all available books and then, using CacheManager
, we populate the book cache.
@Component
public class ApplicationRunner implements CommandLineRunner {
@Autowired
private BookDao dao;
@Autowired
private CacheManager cacheManager;
@Bean
public CacheManager cacheManager() {
return new ConcurrentMapCacheManager("books");
}
@Override
public void run(String... args) throws Exception {
List<Book> results = dao.findAll();
results.forEach(book ->
cacheManager.getCache("books").put(book.getId(), book));
}
}
Solution 4:
If having all instances of Book in memory at startup is your requirement than you should store them in some buffer yourself. Putting them in the cache with the findAll() method means that you must annotate findAll() with @Cacheable. Then you would have to call findAll() at startup. But that does not mean that calling getByIsbn(String isbn) will access the cache even if the corresponding instance has been put in the cache when calling findAll(). Actually it won't because ehcache will cache method return value as a key/value pair where key is computed when method is called. Therefore I don't see how you could match the return value of findAll() and return value of getByIsbn(String) because returned types are not the same and moreover key won't never match for all your instances.
Solution 5:
As Olivier has specified, since spring caches output of function as a single object, using @cacheable notation with findAll will not allow you to load all objects in cache such that they can later be accessed individually.
One possible way you can load all objects in cache is if caching solution being used provides you a way to load all objects at startup. E.g solutions like NCache / TayzGrid provides Cache startup loader feature, that allows you to load cache at startup with objects using a configurable cache startup loader.