Second Level Cache in JPA Explained
8 likes4,162 views
This document discusses second-level caching in Java Persistence API (JPA). It begins with an introduction to the speaker's background and experience. The presentation agenda is then outlined, covering why caching is important, the differences between first-level and second-level caches in JPA, JPA configuration parameters and API for caching, and specifics of Hibernate and EclipseLink second-level caching implementations. Examples are provided that demonstrate performance improvements when utilizing caching strategies like fetching relationships, projections, and aggregation queries.
![Sum
of
Salaries
by
Country
Projection
(3)
Query query = em.createQuery(
"SELECT e.salary, e.address.country
FROM Employee e");
List<Object[]> list = (List<Object[]>) query.getResultList();
// calculate sum of salaries by country
// map: country->sum
Map<String, BigDecimal> results = new HashMap<>();
for (Object[] e : list) {
String country = (String) e[1];
BigDecimal total = results.get(country);
if (total == null) total = BigDecimal.ZERO;
total = total.add((BigDecimal) e[0]);
results.put(country, total);
}
17](https://image.slidesharecdn.com/javaone2016-jpa-2nd-lvl-cache-160920210015/85/Second-Level-Cache-in-JPA-Explained-17-320.jpg)
![Sum
of
Salaries
by
Country
Aggregation
JPQL
(4)
Query query = em.createQuery(
"SELECT SUM(e.salary), e.address.country
FROM Employee e
GROUP BY e.address.country");
List<Object[]> list = (List<Object[]>) query.getResultList();
// already calculated!
18](https://image.slidesharecdn.com/javaone2016-jpa-2nd-lvl-cache-160920210015/85/Second-Level-Cache-in-JPA-Explained-18-320.jpg)
Second Level Cache in JPA Explained
- 1. Second-‐Level Cache in JPA Explained Patrycja Wegrzynowicz CTO, Yonita, Inc. JavaOne 2016
- 2. About Me • 15+ professional experience • SoPware engineer, architect, head of soPware R&D • Author and speaker • JavaOne, Devoxx, JavaZone, TheServerSide Java Symposium, Jazoon, OOPSLA, ASE, others • Top 10 Women in Tech 2016 in Poland • Founder and CTO of Yonita • Automated detecXon and refactoring of soPware defects • Trainings and code reviews • Security, performance, concurrency, databases • TwiYer @yonlabs
- 3. About Me • 15+ professional experience • SoPware engineer, architect, head of soPware R&D • Author and speaker • JavaOne, Devoxx, JavaZone, TheServerSide Java Symposium, Jazoon, OOPSLA, ASE, others • Top 10 Women in Tech 2016 in Poland • Founder and CTO of Yonita • Automated detecXon and refactoring of soPware defects • Trainings and code reviews • Security, performance, concurrency, databases • TwiYer @yonlabs
- 4. Agenda • Why cacheing is important? • 1st Level Cache and 2nd Level Cache • JPA configuraXon parameters for cache • JPA API for cache • Hibernate 2nd Level Cache • EclipseLink 2nd Level Cache (a bit)
- 5. Databases
- 6. Databases
- 7. Performance
- 13. Example
- 14. Employee Entity @Entity public class Employee { @Id @GeneratedValue private Long id; private String firstName; private String lastName; private BigDecimal salary; @OneToOne @JoinColumn(name = "address_id") private Address address; @Temporal(TemporalType.DATE) private Date startDate; @Temporal(TemporalType.DATE) private Date endDate; @ManyToOne @JoinColumn(name = "manager_id") private Employee manager; // … } 14
- 15. Sum of Salaries By Country Select All (1) TypedQuery<Employee> query = em.createQuery( "SELECT e FROM Employee e", Employee.class); List<Employee> list = query.getResultList(); // calculate sum of salaries by country // map: country->sum Map<String, BigDecimal> results = new HashMap<>(); for (Employee e : list) { String country = e.getAddress().getCountry(); BigDecimal total = results.get(country); if (total == null) total = BigDecimal.ZERO; total = total.add(e.getSalary()); results.put(country, total); } 15
- 16. Sum of Salaries by Country Select Join Fetch (2) TypedQuery<Employee> query = em.createQuery( "SELECT e FROM Employee e JOIN FETCH e.address", Employee.class); List<Employee> list = query.getResultList(); // calculate sum of salaries by country // map: country->sum Map<String, BigDecimal> results = new HashMap<>(); for (Employee e : list) { String country = e.getAddress().getCountry(); BigDecimal total = results.get(country); if (total == null) total = BigDecimal.ZERO; total = total.add(e.getSalary()); results.put(country, total); } 16
- 17. Sum of Salaries by Country Projection (3) Query query = em.createQuery( "SELECT e.salary, e.address.country FROM Employee e"); List<Object[]> list = (List<Object[]>) query.getResultList(); // calculate sum of salaries by country // map: country->sum Map<String, BigDecimal> results = new HashMap<>(); for (Object[] e : list) { String country = (String) e[1]; BigDecimal total = results.get(country); if (total == null) total = BigDecimal.ZERO; total = total.add((BigDecimal) e[0]); results.put(country, total); } 17
- 18. Sum of Salaries by Country Aggregation JPQL (4) Query query = em.createQuery( "SELECT SUM(e.salary), e.address.country FROM Employee e GROUP BY e.address.country"); List<Object[]> list = (List<Object[]>) query.getResultList(); // already calculated! 18
- 19. Comparison 1-‐4 (Hibernate) 100000 Employees, Different DB LocaXons Local DB (ping: ~0.05ms) North California (ping: ~38ms) EU Frankfurt (ping: ~420ms) (1) Select All (N+1) 26756ms 2-3 hours ~1 day (2) Select Join Fetch (3) Projection (4) Aggregation JPQL
- 20. Comparison 1-‐4 100000 Employees, Different DB LocaXons Local DB (ping: ~0.05ms) North California (ping: ~38ms) EU Frankfurt (ping: ~420ms) (1) Select All (N+1) 26756ms 2-3 hours ~1 day (2) Select Join Fetch 4854ms 18027ms 25096ms (3) Projection (4) Aggregation JPQL
- 21. Comparison 1-‐4 100000 Employees, Different DB LocaXons Local DB (ping: ~0.05ms) North California (ping: ~38ms) EU Frankfurt (ping: ~420ms) (1) Select All (N+1) 26756ms 2-3 hours ~1 day (2) Select Join Fetch 4854ms 18027ms 25096ms (3) Projection 653ms 2902ms 5006ms (4) Aggregation JPQL
- 22. Comparison 1-‐4 100000 Employees, Different DB LocaXons Local DB (ping: ~0.05ms) North California (ping: ~38ms) EU Frankfurt (ping: ~420ms) (1) Select All (N+1) 26756ms 2-3 hours ~1 day (2) Select Join Fetch 4854ms 18027ms 25096ms (3) Projection 653ms 2902ms 5006ms (4) Aggregation JPQL 182ms 353ms 1198ms
- 23. Performance Tuning: Data • Get your data in bigger chunks • Many small queries => many round-‐trips => huge extra Xme on transport => high latency • Move your data closer to the processing place • Large distance to data => long round-‐trip => high latency • Don’t ask about the same data many Xmes • Extra processing Xme + extra transport Xme Cache
- 24. Cache is Everywhere Web Cache Application Cache RDBMS Cache DNS Cache OS Files Cache
- 25. Second Level Cache in JPA
- 26. JPA Spec • “Persistence providers are not required to support a second-‐level cache.” • “Portable applicaXons should not rely on support by persistence providers for a second-‐level cache.”
- 27. JPA Providers Poll A. Hibernate B. EclipseLink C. OpenJPA D. DataNuclues E. Other
- 28. • Second Level Cache • Persistence Unit • EnXtyManagerFactory • Thread-‐safe, shared • Available for many en1ty managers from one en1ty manager factory • Provider specific support JPA Caches • First Level Cache • Persistence Context • EnXtyManager • Not thread-‐safe • Available for many transac1ons on one en1ty manager • Always
- 30. Puzzle #1 em.getTransaction().begin(); Employee employee = em.find(Employee.class, 2L); employee.getAddress().size(); Employee another = em.find(Employee.class, 2L); em.getTransaction().commit(); (A) No cache used (B) First Level Cache Used (C) Second Level Cache Used (D) None of the above 30
- 31. Loading EntityManager Factory 2LC EntityManager 1LC load from 1LC load from 2LC select from database NOT FOUND NOT FOUND FOUND FOUND RESULT FIND
- 32. Puzzle #1 EntityManager em = emf.createEntityManager(); em.getTransaction().begin(); Employee employee = em.find(Employee.class, 2L); employee.getAddress().size(); Employee another = em.find(Employee.class, 2L); em.getTransaction().commit(); (A) No cache used (B) First Level Cache Used (C) Second Level Cache Used (D) None of the above 32
- 33. Puzzle #1 EntityManager em = emf.createEntityManager(); em.getTransaction().begin(); Employee employee = em.find(Employee.class, 2L); employee.getAddress().size(); Employee another = em.find(Employee.class, 2L); em.getTransaction().commit(); (A) No cache used (B) First Level Cache Used (C) Second Level Cache Used (D) None of the above 33
- 34. Puzzle #2 EntityManager em1 = emf.createEntityManager(); em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); EntityManager em2 = emf.createEntityManager(); em2.getTransaction().begin(); Employee employee = em2.find(Employee.class, 2L); employee.getAddress().size(); em2.getTransaction().commit(); (A) No cache used (B) First Level Cache used (C) Second Level Cache used (D) None of the above 34
- 35. Puzzle #2 EntityManager em1 = emf.createEntityManager(); em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); EntityManager em2 = emf.createEntityManager(); em2.getTransaction().begin(); Employee employee = em2.find(Employee.class, 2L); employee.getAddress().size(); em2.getTransaction().commit(); (A) No cache used (B) First Level Cache used (C) Second Level Cache used (D) None of the above 35
- 36. Puzzle #3 (2LC Configured, Hibernate) EntityManager em1 = emf.createEntityManager(); em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); EntityManager em2 = emf.createEntityManager(); em2.getTransaction().begin(); Employee employee = em2.find(Employee.class, 2L); employee.getAddress().size(); em2.getTransaction().commit(); (A) No cache used (B) First Level Cache used (C) Second Level Cache used (D) None of the above 36
- 37. Puzzle #3 (2LC Configured, Hibernate) EntityManager em1 = emf.createEntityManager(); em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); EntityManager em2 = emf.createEntityManager(); em2.getTransaction().begin(); Employee employee = em2.find(Employee.class, 2L); employee.getAddress().size(); em2.getTransaction().commit(); (A) No cache used (B) First Level Cache used (C) Second Level Cache used (D) None of the above 37
- 38. Puzzle #4 (2LC Configured!) em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); a em2.getTransaction().begin(); TypedQuery<Employee> q = em.createQuery("SELECT e FROM Employee e WHERE e.id=:id", Employee.class); q.setParameter("id", 2l); Employee employee = q.getSingleResult(); employee.getAddress().size(); em2.getTransaction().commit(); (A) No cache used (B) First Level Cache used (C) Second Level Cache used (D) None of the above 38
- 39. Puzzle #4 (2LC Configured!) em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); a em2.getTransaction().begin(); TypedQuery<Employee> q = em.createQuery("SELECT e FROM Employee e WHERE e.id=:id", Employee.class); q.setParameter("id", 2l); Employee employee = q.getSingleResult(); employee.getAddress().size(); em2.getTransaction().commit(); (A) No cache used (B) First Level Cache used (C) Second Level Cache used (D) None of the above 39
- 41. Programmatic Properties props = new Properties() .add(“javax.persistence.sharedCache.mode”, “ENABLE_SELECTIVE”); EntityManagerFactor emf = Persistence .createEntityManagerFactory(“test-pu”, props); 41
- 42. JPA Cache Modes • ALL • All enXty data is stored in the second-‐level cache for this persistence unit. • NONE • No data is cached in the persistence unit. The persistence provider must not cache any data. • ENABLE_SELECTIVE • Enable caching for enXXes that have been explicitly set with the @Cacheable annotaXon. • DISABLE_SELECTIVE • Enable caching for all enXXes except those that have been explicitly set with the @Cacheable(false) annotaXon. • UNSPECIFIED • The caching behavior for the persistence unit is undefined. The persistence provider’s default caching behavior will be used.
- 43. JPA Cache Modes • ALL • All enXty data is stored in the second-‐level cache for this persistence unit. • NONE • No data is cached in the persistence unit. The persistence provider must not cache any data. • ENABLE_SELECTIVE • Enable caching for enXXes that have been explicitly set with the @Cacheable annotaXon. • DISABLE_SELECTIVE • Enable caching for all enXXes except those that have been explicitly set with the @Cacheable(false) annotaXon. • UNSPECIFIED • The caching behavior for the persistence unit is undefined. The persistence provider’s default caching behavior will be used.
- 44. @Cachable @Entity @Cacheable public class Employee { } @Entity @Cacheable(true) public class Employee { } @Entity @Cacheable(false) public class Address { } 44
- 45. @Cacheable • @Cacheable ignored for ALL or NONE
- 46. Cache Retrieval and Store Modes • Cache Retrieval Modes • javax.persistence.CacheRetrieveMode • USE (default) • BYPASS • Cache Store Modes • javax.persistence.storeMode • USE (default) • the cache data is created or updated when data is read from or commiYed to • when data is already in the cache, no refresh on read • REFRESH • forced refresh on read • BYPASS
- 47. Cache Retrieval and Store EntityManager em = ...; em.setProperty("javax.persistence.cache.storeMode", “BYPASS"); Map<String, Object> props = new HashMap<String, Object>(); props.put("javax.persistence.cache.retrieveMode", "BYPASS"); Employee employee = = em.find(Employee.class, 1L, props); TypedQuery<Employee> q = em.createQuery(cq); q.setHint("javax.persistence.cache.storeMode", "REFRESH"); 47
- 48. Programmatic Access to Cache EntityManager em = ...; Cache cache = em.getEntityManagerFactory().getCache(); if (cache.contains(Employee.class, 1L)) { // the data is cached } else { // the data is NOT cached } Cache interface methods: boolean contains(Class cls, Object primaryKey) void evict(Class cls) void evict(Class cls, Object primaryKey) void evictAll() <T> T unwrap(Class<T> cls) 48
- 49. Forget that! • Don’t use cache retrieval and store modes • Don’t use programmaXc access to 2LC • Use provider-‐specific configuraXon!
- 50. Hibernate Caches
- 51. Hibernate Caches • First Level Cache • Second Level Cache • hydrated or disassembled enXXes: EnXtyEntry • collecXons • Query Cache
- 52. Hibernate Cache ConfiguraXon • hibernate.cache.use_second_level_cache • Enable or disable second level caching overall. • Default is true • hibernate.cache.region.factory_class • Default region factory is NoCachingRegionFactory • hibernate.cache.use_query_cache • Enable or disable second level caching of query results. • Default is false. • hibernate.cache.query_cache_factory
- 53. Hibernate Cache ConfiguraXon • hibernate.cache.use_minimal_puts • OpXmizes second-‐level cache operaXons to minimize writes, at the cost of more frequent reads. Providers typically set this appropriately. • hibernate.cache.default_cache_concurrency_strategy • In Hibernate second-‐level caching, all regions can be configured differently including the concurrency strategy to use when accessing that parXcular region. This sevng allows to define a default strategy to be used. • Providers specify this sevng!
- 54. Hibernate Cache ConfiguraXon • hibernate.cache.use_structured_entries • If true, forces Hibernate to store data in the second-‐level cache in a more human-‐friendly format. • Default: false • hibernate.cache.auto_evict_collecXon_cache • Enables or disables the automaXc evicXon of a bidirecXonal associaXon’s collecXon cache entry when the associaXon is changed just from the owning side. • Default: false • hibernate.cache.use_reference_entries • Enable direct storage of enXty references into the second level cache for read-‐only or immutable enXXes.
- 55. Cache Concurrency Strategy • Global cache concurrency strategy • hibernate.cache.default_cache_concurrency_strategy • Hibernate @Cache annotaXon on an enXty level • usage: defines the CacheConcurrencyStrategy • region: defines a cache region where entries will be stored • include: if lazy properXes should be included in the second level cache. Default value is "all", so lazy properXes are cacheable. The other possible value is "non-‐lazy", so lazy properXes are not cacheable.
- 56. Cache Concurrency Strategies • read-‐only • ApplicaXon read-‐only data • Allows deletes • read-‐write • ApplicaXon updates data • Consistent access to a single enXty, but not a serializable transacXon isolaXon level • nonstrict-‐read-‐write • Occasional stale reads • transacXonal • Provides serializable transacXon isolaXon level
- 57. Example – Entity and Collection Cache@Entity @Cacheable @org.hibernate.annotations.Cache(usage = CacheConcurrencyStrategy.READ_WRITE) public class Employee { @OneToMany(mappedBy = "employee", cascade = CascadeType.ALL) @org.hibernate.annotations.Cache(usage = CacheConcurrencyStrategy.NONSTRICT_READ_WRITE) private Set<Phone> phones = new HashSet<>(); } // cache used! Person person = entityManager.find(Employee.class, 1L); // cache used! Person person = entityManager.find(Employee.class, 1L); person.getPhones().size(); 57
- 58. Example -‐ Query Cache List<Employee> employees = entityManager.createQuery( "select e " + "from Employee e " + "where e.firstName = :firstName", Employee.class) .setParameter( "firstName", "John") .setHint("org.hibernate.cacheable", "true") .getResultList(); 58
- 59. Puzzle #2 EntityManager em1 = emf.createEntityManager(); em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); EntityManager em2 = emf.createEntityManager(); em2.getTransaction().begin(); Employee employee = em2.find(Employee.class, 2L); employee.getAddress().size(); em2.getTransaction().commit(); 59
- 60. find, no collecXon caching Local DB (ping: ~0.05ms) North California (ping: ~38ms) EU Frankfurt (ping: ~420ms) No cache 87ms 186ms 1164ms Cached 62ms 127ms 995ms
- 61. find, collecXon caching Local DB (ping: ~0.05ms) North California (ping: ~38ms) EU Frankfurt (ping: ~420ms) No cache 82ms 162ms 1178ms Cached 3ms 98ms 941ms
- 62. Puzzle #2/P6Spy EntityManager em1 = emf.createEntityManager(); em1.getTransaction().begin(); Employee employee = em1.find(Employee.class, 2L); employee.getAddress().size(); em1.getTransaction().commit(); EntityManager em2 = emf.createEntityManager(); em2.getTransaction().begin(); Employee employee = em2.find(Employee.class, 2L); employee.getAddress().size(); em2.getTransaction().commit(); 62
- 63. Available Hibernate 2nd Level Cache ImplementaXons • EHCache • Infinispan • Hazelcast
- 64. Infinispan Configuration (Local) <!-- This configuration is suitable for non-clustered environments, where only single instance accesses the DB --> <cache-container name="SampleCacheManager" statistics="false" default-cache="the-default-cache" shutdown-hook="DEFAULT"> <jmx duplicate-domains="true"/> <local-cache-configuration name="the-default-cache" statistics="false" /> <!-- Default configuration is appropriate for entity/collection caching. --> <local-cache-configuration name="entity" simple-cache="true" statistics="false" statistics-available="false"> <transaction mode="NONE" /> <eviction max-entries="10000" strategy="LRU"/> <expiration max-idle="100000" interval="5000"/> </local-cache-configuration> <!-- A config appropriate for query caching. Does not replicate queries. --> <local-cache-configuration name="local-query" simple-cache="true" statistics="false" statistics-available="false"> <transaction mode="NONE" /> <eviction max-entries="10000" strategy="LRU"/> <expiration max-idle="100000" interval="5000"/> </local-cache-configuration> <local-cache-configuration name="timestamps" simple-cache="true" statistics="false" statistics-available="false"> <locking concurrency-level="1000" acquire-timeout="15000"/> <!-- Explicitly non transactional --> <transaction mode="NONE"/> <!-- Don't ever evict modification timestamps --> <eviction strategy="NONE"/> <expiration interval="0"/> </local-cache-configuration> <!-- When providing custom configuration, always make this cache local and non-transactional. 64
- 65. Infinispan Configuration (Clustered)<jgroups> <stack-file name="hibernate-jgroups" path="${hibernate.cache.infinispan.jgroups_cfg:default-configs/default-jgroups-tcp.xml}"/> </jgroups> <cache-container name="SampleCacheManager" statistics="false" default-cache="the-default-cache" shutdown-hook="DEFAULT"> <transport stack="hibernate-jgroups" cluster="infinispan-hibernate-cluster"/> <jmx duplicate-domains="true"/> <local-cache-configuration name="the-default-cache" statistics="false" /> <!-- Default configuration is appropriate for entity/collection caching. --> <invalidation-cache-configuration name="entity" mode="SYNC" remote-timeout="20000" statistics="false" statistics-available="false"> <locking concurrency-level="1000" acquire-timeout="15000"/> <transaction mode="NONE" /> <eviction max-entries="10000" strategy="LRU"/> <expiration max-idle="100000" interval="5000"/> </invalidation-cache-configuration> <!-- A config appropriate for query caching. Does not replicate queries. --> <local-cache-configuration name="local-query" statistics="false" statistics-available="false"> <locking concurrency-level="1000" acquire-timeout="15000"/> <transaction mode="NONE" /> <eviction max-entries="10000" strategy="LRU"/> <expiration max-idle="100000" interval="5000"/> </local-cache-configuration> <!-- A query cache that replicates queries. Replication is asynchronous. --> <replicated-cache-configuration name="replicated-query" mode="ASYNC" statistics="false" statistics-available="false"> <locking concurrency-level="1000" acquire-timeout="15000"/> <transaction mode="NONE" /> <eviction max-entries="10000" strategy="LRU"/> <expiration max-idle="100000" interval="5000"/> </replicated-cache-configuration> <!-- Optimized for timestamp caching. A clustered timestamp cache is required if query caching is used, even if the query cache itself is configured with CacheMode=LOCAL. --> 65
- 66. Guidelines • Cache as much as you can • As much RAM you have • Do it wisely • Read-‐only data • Almost read-‐only data • More reads than writes • Hit raXo
- 67. EclipseLink • 2LC enabled by default • CollecXon caching • Query caching • @org.eclipse.persistence.annotaXons.Cache • type: type of the cache (FULL, WEAK, SOFT, SOFT_WEAK, HARD_WEAK) • size: number of objects • isolaXon: shared, isolated, protected • expiry • expiryTimeOfDay • alwaysRefresh • refreshOnlyIfNewer • disableHits • coordinaXonType • databaseChangeNoXficaXonType
- 68. Conclusion
- 69. A fool with a tool is only a fool!
- 71. Please, vote! :)
- 72. Q&A • patrycja@yonita.com • @yonlabs