ShedLock makes sure that your scheduled tasks are executed at most once at the same time. If a task is being executed on one node, it acquires a lock which prevents execution of the same task from another node (or thread). Please note, that if one task is already being executed on one node, execution on other nodes does not wait, it is simply skipped.
ShedLock uses an external store like Mongo, JDBC database, Redis, Hazelcast, ZooKeeper or others for coordination.
Feedback and pull-requests welcome!
Please note that ShedLock is not and will never be full-fledged scheduler, it's just a lock. If you need a distributed scheduler, please use another project (db-scheduler, JobRunr). ShedLock is designed to be used in situations where you have scheduled tasks that are not ready to be executed in parallel, but can be safely executed repeatedly. Moreover, the locks are time-based and ShedLock assumes that clocks on the nodes are synchronized.
- Versions
- Components
- Usage
- Lock Providers
- JdbcTemplate
- R2DBC
- jOOQ
- Micronaut Data Jdbc
- Mongo
- DynamoDB
- DynamoDB 2
- ZooKeeper (using Curator)
- Redis (using Spring RedisConnectionFactory)
- Redis (using Spring ReactiveRedisConnectionFactory)
- Redis (using Jedis)
- Hazelcast
- Couchbase
- ElasticSearch
- OpenSearch
- CosmosDB
- Cassandra
- Consul
- ArangoDB
- Neo4j
- Etcd
- Apache Ignite
- In-Memory
- Memcached
- Datastore
- Multi-tenancy
- Customization
- Duration specification
- Extending the lock
- Micronaut integration
- CDI integration
- Locking without a framework
- Troubleshooting
- Modes of Spring integration
- Release notes
If you are using JDK >17 and up-to-date libraries like Spring 6, use version 5.1.0 (Release Notes). If you are on older JDK or library, use version 4.44.0 (documentation).
Shedlock consists of three parts
- Core - The locking mechanism
- Integration - integration with your application, using Spring AOP, Micronaut AOP or manual code
- Lock provider - provides the lock using an external process like SQL database, Mongo, Redis and others
To use ShedLock, you do the following
- Enable and configure Scheduled locking
- Annotate your scheduled tasks
- Configure a Lock Provider
First of all, we have to import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-spring</artifactId>
<version>5.7.0</version>
</dependency>
Now we need to integrate the library with Spring. In order to enable schedule locking use @EnableSchedulerLock
annotation
@Configuration
@EnableScheduling
@EnableSchedulerLock(defaultLockAtMostFor = "10m")
class MySpringConfiguration {
...
}
import net.javacrumbs.shedlock.spring.annotation.SchedulerLock;
...
@Scheduled(...)
@SchedulerLock(name = "scheduledTaskName")
public void scheduledTask() {
// To assert that the lock is held (prevents misconfiguration errors)
LockAssert.assertLocked();
// do something
}
The @SchedulerLock
annotation has several purposes. First of all, only annotated methods are locked, the library ignores
all other scheduled tasks. You also have to specify the name for the lock. Only one task with the same name can be executed
at the same time.
You can also set lockAtMostFor
attribute which specifies how long the lock should be kept in case the
executing node dies. This is just a fallback, under normal circumstances the lock is released as soon the tasks finishes.
You have to set lockAtMostFor
to a value which is much longer than normal execution time. If the task takes longer than
lockAtMostFor
the resulting behavior may be unpredictable (more than one process will effectively hold the lock).
If you do not specify lockAtMostFor
in @SchedulerLock
default value from @EnableSchedulerLock
will be used.
Lastly, you can set lockAtLeastFor
attribute which specifies minimum amount of time for which the lock should be kept.
Its main purpose is to prevent execution from multiple nodes in case of really short tasks and clock difference between the nodes.
All the annotations support Spring Expression Language (SpEL).
Let's say you have a task which you execute every 15 minutes and which usually takes few minutes to run. Moreover, you want to execute it at most once per 15 minutes. In that case, you can configure it like this:
import net.javacrumbs.shedlock.core.SchedulerLock;
@Scheduled(cron = "0 */15 * * * *")
@SchedulerLock(name = "scheduledTaskName", lockAtMostFor = "14m", lockAtLeastFor = "14m")
public void scheduledTask() {
// do something
}
By setting lockAtMostFor
we make sure that the lock is released even if the node dies and by setting lockAtLeastFor
we make sure it's not executed more than once in fifteen minutes.
Please note that lockAtMostFor
is just a safety net in case that the node executing the task dies, so set it to
a time that is significantly larger than maximum estimated execution time. If the task takes longer than lockAtMostFor
,
it may be executed again and the results will be unpredictable (more processes will hold the lock).
There are several implementations of LockProvider.
First, create lock table (please note that name
has to be primary key)
# MySQL, MariaDB
CREATE TABLE shedlock(name VARCHAR(64) NOT NULL, lock_until TIMESTAMP(3) NOT NULL,
locked_at TIMESTAMP(3) NOT NULL DEFAULT CURRENT_TIMESTAMP(3), locked_by VARCHAR(255) NOT NULL, PRIMARY KEY (name));
# Postgres
CREATE TABLE shedlock(name VARCHAR(64) NOT NULL, lock_until TIMESTAMP NOT NULL,
locked_at TIMESTAMP NOT NULL, locked_by VARCHAR(255) NOT NULL, PRIMARY KEY (name));
# Oracle
CREATE TABLE shedlock(name VARCHAR(64) NOT NULL, lock_until TIMESTAMP(3) NOT NULL,
locked_at TIMESTAMP(3) NOT NULL, locked_by VARCHAR(255) NOT NULL, PRIMARY KEY (name));
# MS SQL
CREATE TABLE shedlock(name VARCHAR(64) NOT NULL, lock_until datetime2 NOT NULL,
locked_at datetime2 NOT NULL, locked_by VARCHAR(255) NOT NULL, PRIMARY KEY (name));
# DB2
CREATE TABLE shedlock(name VARCHAR(64) NOT NULL PRIMARY KEY, lock_until TIMESTAMP NOT NULL,
locked_at TIMESTAMP NOT NULL, locked_by VARCHAR(255) NOT NULL);
Or use this liquibase change-set.
Add dependency
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-jdbc-template</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.jdbctemplate.JdbcTemplateLockProvider;
...
@Bean
public LockProvider lockProvider(DataSource dataSource) {
return new JdbcTemplateLockProvider(
JdbcTemplateLockProvider.Configuration.builder()
.withJdbcTemplate(new JdbcTemplate(dataSource))
.usingDbTime() // Works on Postgres, MySQL, MariaDb, MS SQL, Oracle, DB2, HSQL and H2
.build()
);
}
By specifying usingDbTime()
the lock provider will use UTC time based on the DB server clock.
If you do not specify this option, clock from the app server will be used (the clocks on app servers may not be
synchronized thus leading to various locking issues).
It's strongly recommended to use usingDbTime()
option as it uses DB engine specific SQL that prevents INSERT conflicts.
See more details here.
For more fine-grained configuration use other options of the Configuration
object
new JdbcTemplateLockProvider(builder()
.withTableName("shdlck")
.withColumnNames(new ColumnNames("n", "lck_untl", "lckd_at", "lckd_by"))
.withJdbcTemplate(new JdbcTemplate(getDatasource()))
.withLockedByValue("my-value")
.withDbUpperCase(true)
.build())
If you need to specify a schema, you can set it in the table name using the usual dot notation
new JdbcTemplateLockProvider(datasource, "my_schema.shedlock")
To use a database with case-sensitive table and column names, the .withDbUpperCase(true)
flag can be used.
Default is false
(lowercase).
Do not manually delete lock row from the DB table. ShedLock has an in-memory cache of existing lock rows so the row will NOT be automatically recreated until application restart. If you need to, you can edit the row/document, risking only that multiple locks will be held.
If you are really brave, you can try experimental R2DBC support. Please keep in mind that the capabilities of this lock provider are really limited and that the whole ecosystem around R2DBC is in flux and may easily break.
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-r2dbc</artifactId>
<version>5.7.0</version>
</dependency>
and use it.
@Override
protected LockProvider getLockProvider() {
return new R2dbcLockProvider(connectionFactory);
}
I recommend using R2DBC connection pool.
First, create lock table as described in the JdbcTemplate section above.
Add dependency
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-jooq</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.jooq;
...
@Bean
public LockProvider getLockProvider(DSLContext dslContext) {
return new JooqLockProvider(dslContext);
}
jOOQ provider has a bit different transactional behavior. While the other JDBC lock providers create new transaction (with REQUIRES_NEW), jOOQ does not support setting it. ShedLock tries to create a new transaction, but depending on your set-up, ShedLock DB operations may end-up being part of the enclosing transaction.
jOOQ lock provider is not yet as flexible as its JdbcTemplate counterpart, but it can be. If you need some configuration option, please let me know.
If you are using Micronaut data and you do not want to add dependency on Spring JDBC, you can use Micronaut JDBC support. Just be aware that it has just a basic functionality when compared to the JdbcTemplate provider.
First, create lock table as described in the JdbcTemplate section above.
Add dependency
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-jdbc-micronaut</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.jdbc.micronaut.MicronautJdbcLockProvider;
...
@Singleton
public LockProvider lockProvider(TransactionOperations<Connection> transactionManager) {
return new MicronautJdbcLockProvider(transactionManager);
}
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-mongo</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.mongo.MongoLockProvider;
...
@Bean
public LockProvider lockProvider(MongoClient mongo) {
return new MongoLockProvider(mongo.getDatabase(databaseName))
}
Please note that MongoDB integration requires Mongo >= 2.4 and mongo-java-driver >= 3.7.0
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-mongo-reactivestreams</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.mongo.reactivestreams.ReactiveStreamsMongoLockProvider;
...
@Bean
public LockProvider lockProvider(MongoClient mongo) {
return new ReactiveStreamsMongoLockProvider(mongo.getDatabase(databaseName))
}
Please note that MongoDB integration requires Mongo >= 4.x and mongodb-driver-reactivestreams 1.x
Depends on AWS SDK v2.
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-dynamodb2</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.dynamodb2.DynamoDBLockProvider;
...
@Bean
public LockProvider lockProvider(software.amazon.awssdk.services.dynamodb.DynamoDbClient dynamoDB) {
return new DynamoDBLockProvider(dynamoDB, "Shedlock");
}
Please note that the lock table must be created externally with
_id
as a partition key.DynamoDBUtils#createLockTable
may be used for creating it programmatically. A table definition is available fromDynamoDBLockProvider
's Javadoc.
Import
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-zookeeper-curator</artifactId>
<version>5.7.0</version>
</dependency>
and configure
import net.javacrumbs.shedlock.provider.zookeeper.curator.ZookeeperCuratorLockProvider;
...
@Bean
public LockProvider lockProvider(org.apache.curator.framework.CuratorFramework client) {
return new ZookeeperCuratorLockProvider(client);
}
By default, nodes for locks will be created under /shedlock
node.
Import
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-redis-spring</artifactId>
<version>5.7.0</version>
</dependency>
and configure
import net.javacrumbs.shedlock.provider.redis.spring.RedisLockProvider;
import org.springframework.data.redis.connection.RedisConnectionFactory;
...
@Bean
public LockProvider lockProvider(RedisConnectionFactory connectionFactory) {
return new RedisLockProvider(connectionFactory, ENV);
}
Import
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-redis-spring</artifactId>
<version>5.7.0</version>
</dependency>
and configure
import net.javacrumbs.shedlock.provider.redis.spring.ReactiveRedisLockProvider;
import org.springframework.data.redis.connection.ReactiveRedisConnectionFactory;
...
@Bean
public LockProvider lockProvider(ReactiveRedisConnectionFactory connectionFactory) {
return new ReactiveRedisLockProvider.Builder(connectionFactory)
.environment(ENV)
.build();
}
Redis lock provider uses classical lock mechanism as described here which may not be reliable in case of Redis master failure.
Import
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-redis-jedis4</artifactId>
<version>5.7.0</version>
</dependency>
and configure
import net.javacrumbs.shedlock.provider.redis.jedis.JedisLockProvider;
...
@Bean
public LockProvider lockProvider(JedisPool jedisPool) {
return new JedisLockProvider(jedisPool, ENV);
}
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-hazelcast4</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.hazelcast4.HazelcastLockProvider;
...
@Bean
public HazelcastLockProvider lockProvider(HazelcastInstance hazelcastInstance) {
return new HazelcastLockProvider(hazelcastInstance);
}
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-couchbase-javaclient3</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.couchbase.javaclient.CouchbaseLockProvider;
...
@Bean
public CouchbaseLockProvider lockProvider(Bucket bucket) {
return new CouchbaseLockProvider(bucket);
}
For Couchbase 3 use shedlock-provider-couchbase-javaclient3
module and net.javacrumbs.shedlock.provider.couchbase3
package.
I am really not sure it's a good idea to use Elasticsearch as a lock provider. But if you have no other choice, you can. Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-elasticsearch8</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import static net.javacrumbs.shedlock.provider.elasticsearch8.ElasticsearchLockProvider;
...
@Bean
public ElasticsearchLockProvider lockProvider(ElasticsearchClient client) {
return new ElasticsearchLockProvider(client);
}
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-opensearch</artifactId>
<version>4.36.1</version>
</dependency>
Configure:
import static net.javacrumbs.shedlock.provider.opensearch.OpenSearchLockProvider;
...
@Bean
public OpenSearchLockProvider lockProvider(RestHighLevelClient highLevelClient) {
return new OpenSearchLockProvider(highLevelClient);
}
CosmosDB support is provided by a third-party module available here
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-cassandra</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.cassandra.CassandraLockProvider;
import net.javacrumbs.shedlock.provider.cassandra.CassandraLockProvider.Configuration;
...
@Bean
public CassandraLockProvider lockProvider(CqlSession cqlSession) {
return new CassandraLockProvider(Configuration.builder().withCqlSession(cqlSession).withTableName("lock").build());
}
Example for creating default keyspace and table in local Cassandra instance:
CREATE KEYSPACE shedlock with replication={'class':'SimpleStrategy', 'replication_factor':1} and durable_writes=true;
CREATE TABLE shedlock.lock (name text PRIMARY KEY, lockUntil timestamp, lockedAt timestamp, lockedBy text);
Please, note that CassandraLockProvider uses Cassandra driver v4, which is part of Spring Boot since 2.3.
ConsulLockProvider has one limitation: lockAtMostFor setting will have a minimum value of 10 seconds. It is dictated by consul's session limitations.
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-consul</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.consul.ConsulLockProvider;
...
@Bean // for micronaut please define preDestroy property @Bean(preDestroy="close")
public ConsulLockProvider lockProvider(com.ecwid.consul.v1.ConsulClient consulClient) {
return new ConsulLockProvider(consulClient);
}
Please, note that Consul lock provider uses ecwid consul-api client, which is part of spring cloud consul integration (the spring-cloud-starter-consul-discovery
package).
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-arangodb</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.arangodb.ArangoLockProvider;
...
@Bean
public ArangoLockProvider lockProvider(final ArangoOperations arangoTemplate) {
return new ArangoLockProvider(arangoTemplate.driver().db(DB_NAME));
}
Please, note that ArangoDB lock provider uses ArangoDB driver v6.7, which is part of arango-spring-data in version 3.3.0.
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-neo4j</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.core.LockConfiguration;
...
@Bean
Neo4jLockProvider lockProvider(org.neo4j.driver.Driver driver) {
return new Neo4jLockProvider(driver);
}
Please make sure that neo4j-java-driver
version used by shedlock-provider-neo4j
matches the driver version used in your
project (if you use spring-boot-starter-data-neo4j
, it is probably provided transitively).
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-etcd-jetcd</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.etcd.jetcd.EtcdLockProvider;
...
@Bean
public LockProvider lockProvider(Client client) {
return new EtcdLockProvider(client);
}
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-ignite</artifactId>
<version>5.7.0</version>
</dependency>
Configure:
import net.javacrumbs.shedlock.provider.ignite.IgniteLockProvider;
...
@Bean
public LockProvider lockProvider(Ignite ignite) {
return new IgniteLockProvider(ignite);
}
If you want to use a lock provider in tests there is an in-Memory implementation.
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-inmemory</artifactId>
<version>5.7.0</version>
<scope>test</scope>
</dependency>
import net.javacrumbs.shedlock.provider.inmemory.InMemoryLockProvider;
...
@Bean
public LockProvider lockProvider() {
return new InMemoryLockProvider();
}
Please, be aware that memcached is not a database but a cache. It means that if the cache is full, the lock may be released prematurely Use only if you know what you are doing.
Import
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-memcached-spy</artifactId>
<version>5.7.0</version>
</dependency>
and configure
import net.javacrumbs.shedlock.provider.memcached.spy.MemcachedLockProvider;
...
@Bean
public LockProvider lockProvider(net.spy.memcached.MemcachedClient client) {
return new MemcachedLockProvider(client, ENV);
}
P.S.:
Memcached Standard Protocol:
- A key (arbitrary string up to 250 bytes in length. No space or newlines for ASCII mode)
- An expiration time, in
seconds
. '0' means never expire. Can be up to 30 days. After 30 days, is treated as a unix timestamp of an exact date. (supportseconds
、minutes
、days
, and less than30
days)
Import the project
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-provider-datastore</artifactId>
<version>5.7.0</version>
</dependency>
and configure
import net.javacrumbs.shedlock.provider.datastore.DatastoreLockProvider;
...
@Bean
public LockProvider lockProvider(com.google.cloud.datastore.Datastore datastore) {
return new DatastoreLockProvider(datastore);
}
If you have multi-tenancy use-case you can use a lock provider similar to this one (see the full example)
private static abstract class MultiTenancyLockProvider implements LockProvider {
private final ConcurrentHashMap<String, LockProvider> providers = new ConcurrentHashMap<>();
@Override
public @NonNull Optional<SimpleLock> lock(@NonNull LockConfiguration lockConfiguration) {
String tenantName = getTenantName(lockConfiguration);
return providers.computeIfAbsent(tenantName, this::createLockProvider).lock(lockConfiguration);
}
protected abstract LockProvider createLockProvider(String tenantName) ;
protected abstract String getTenantName(LockConfiguration lockConfiguration);
}
You can customize the behavior of the library by implementing LockProvider
interface. Let's say you want to implement
a special behavior after a lock is obtained. You can do it like this:
public class MyLockProvider implements LockProvider {
private final LockProvider delegate;
public MyLockProvider(LockProvider delegate) {
this.delegate = delegate;
}
@Override
public Optional<SimpleLock> lock(LockConfiguration lockConfiguration) {
Optional<SimpleLock> lock = delegate.lock(lockConfiguration);
if (lock.isPresent()) {
// do something
}
return lock;
}
}
All the annotations where you need to specify a duration support the following formats
- duration+unit -
1s
,5ms
,5m
,1d
(Since 4.0.0) - duration in ms -
100
(only Spring integration) - ISO-8601 -
PT15M
(see Duration.parse() documentation)
There are some use-cases which require to extend currently held lock. You can use LockExtender in the following way:
LockExtender.extendActiveLock(Duration.ofMinutes(5), ZERO);
Please note that not all lock provider implementations support lock extension.
There is also KeepAliveLockProvider that is able to keep the lock alive by periodically extending it. It can be used by wrapping the original lock provider. My personal opinion is that it should be used only in special cases, it adds more complexity to the library and the flow is harder to reason about so please use moderately.
@Bean
public LockProvider lockProvider(...) {
return new KeepAliveLockProvider(new XyzProvider(...), scheduler);
}
KeepAliveLockProvider extends the lock in the middle of the lockAtMostFor interval. For example, if the lockAtMostFor is 10 minutes the lock is extended every 5 minutes for 10 minutes until the lock is released. Please note that the minimal lockAtMostFor time supported by this provider is 30s.
Since version 4.0.0, it's possible to use Micronaut framework for integration
Import the project:
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<artifactId>shedlock-micronaut</artifactId>
<version>5.7.0</version>
</dependency>
Configure default lockAtMostFor value (application.yml):
shedlock:
defaults:
lock-at-most-for: 1m
Configure lock provider:
@Singleton
public LockProvider lockProvider() {
... select and configure your lock provider
}
Configure the scheduled task:
@Scheduled(fixedDelay = "1s")
@SchedulerLock(name = "myTask")
public void myTask() {
assertLocked();
...
}
Since version 5.0.0, it's possible to use CDI for integration (tested only with Quarkus)
Import the project:
<dependency>
<groupId>net.javacrumbs.shedlock</groupId>
<!-- use shedlock-cdi-vintage for quarkus 2.x -->
<artifactId>shedlock-cdi</artifactId>
<version>5.7.0</version>
</dependency>
Configure default lockAtMostFor value (application.properties):
shedlock.defaults.lock-at-most-for=PT30S
Configure lock provider:
@Produces
@Singleton
public LockProvider lockProvider() {
...
}
Configure the scheduled task:
@Scheduled(every = "1s")
@SchedulerLock(name = "myTask")
public void myTask() {
assertLocked();
...
}
The implementation only depends on jakarta.enterprise.cdi-api
and microprofile-config-api
so it should be
usable in other CDI compatible frameworks, but it has not been tested with anything else than Quarkus. It's
built on top of javax annotation as Quarkus has not moved to Jakarta EE namespace yet.
The support is minimalistic, for example there is no support for expressions in the annotation parameters yet, if you need it, feel free to send a PR.
It is possible to use ShedLock without a framework
LockingTaskExecutor executor = new DefaultLockingTaskExecutor(lockProvider);
...
Instant lockAtMostUntil = Instant.now().plusSeconds(600);
executor.executeWithLock(runnable, new LockConfiguration("lockName", lockAtMostUntil));
Some lock providers support extension of the lock. For the time being, it requires manual lock manipulation,
directly using LockProvider
and calling extend
method on the SimpleLock
.
ShedLock supports two modes of Spring integration. One that uses an AOP proxy around scheduled method (PROXY_METHOD) and one that proxies TaskScheduler (PROXY_SCHEDULER)
Since version 4.0.0, the default mode of Spring integration is an AOP proxy around the annotated method.
The main advantage of this mode is that it plays well with other frameworks that want to somehow alter the default Spring scheduling mechanism. The disadvantage is that the lock is applied even if you call the method directly. If the method returns a value and the lock is held by another process, null or an empty Optional will be returned (primitive return types are not supported).
Final and non-public methods are not proxied so either you have to make your scheduled methods public and non-final or use TaskScheduler proxy.
This mode wraps Spring TaskScheduler
in an AOP proxy. This mode does not play well with instrumentation libraries
like opentelementry that also wrap TaskScheduler. Please only use it if you know what you are doing.
It can be switched-on like this (PROXY_SCHEDULER was the default method before 4.0.0):
@EnableSchedulerLock(interceptMode = PROXY_SCHEDULER)
If you do not specify your task scheduler, a default one is created for you. If you have special needs, just create a bean implementing TaskScheduler
interface and it will get wrapped into the AOP proxy automatically.
@Bean
public TaskScheduler taskScheduler() {
return new MySpecialTaskScheduler();
}
Alternatively, you can define a bean of type ScheduledExecutorService
and it will automatically get used by the tasks
scheduling mechanism.
Spring XML configuration is not supported as of version 3.0.0. If you need it, please use version 2.6.0 or file an issue explaining why it is needed.
To prevent misconfiguration errors, like AOP misconfiguration, missing annotation etc., you can assert that the lock works by using LockAssert:
@Scheduled(...)
@SchedulerLock(..)
public void scheduledTask() {
// To assert that the lock is held (prevents misconfiguration errors)
LockAssert.assertLocked();
// do something
}
In unit tests you can switch-off the assertion by calling LockAssert.TestHelper.makeAllAssertsPass(true)
on given thread (as in this example).
The library is tested with Kotlin and works fine. The only issue is Spring AOP which does not work on final method. If you use @SchedulerLock
with @Component
annotation, everything should work since Kotlin Spring compiler plugin will automatically 'open' the method for you. If @Component
annotation is not present, you
have to open the method by yourself. (see this issue for more details)
Locks in ShedLock have an expiration time which leads to the following possible issues.
- If the task runs longer than
lockAtMostFor
, the task can be executed more than once - If the clock difference between two nodes is more than
lockAtLeastFor
or minimal execution time the task can be executed more than once.
Help, ShedLock does not do what it's supposed to do!
- Upgrade to the newest version
- Use LockAssert to ensure that AOP is correctly configured.
- If it does not work, please read about Spring AOP internals (for example here)
- Check the storage. If you are using JDBC, check the ShedLock table. If it's empty, ShedLock is not properly configured. If there is more than one record with the same name, you are missing a primary key.
- Use ShedLock debug log. ShedLock logs interesting information on DEBUG level with logger name
net.javacrumbs.shedlock
. It should help you to see what's going on. - For short-running tasks consider using
lockAtLeastFor
. If the tasks are short-running, they could be executed one after another,lockAtLeastFor
can prevent it.
- JedisLockProvider supports extending (thanks @shotmk)
- Better behavior when locks are nested #1493
- JedisLockProvider (version 3) supports extending (thanks @shotmk)
- JedisLockProvider supports extending (thanks @shotmk)
- Ability to explicitly set database product in JdbTemplateLockProvider (thanks @metron2)
- Removed forgotten versions from BOM
- Dependency updates
- Datastore support (thanks @mmastika)
- Dependency updates
- Handle uncategorized SQL exceptions (thanks @jaam)
- Dependency updates
- Added shedlock-cdi module (supports newest CDI version)
- Dependency updates
- Uppercase in JdbcTemplateProvider (thanks @Ragin-LundF)
- Dependency updates
- Added SpEL support to @SchedulerLock name attribute (thanks @ipalbeniz)
- Dependency updates
- Work around broken Spring 6 exception translation lukas-krecan#1272
- Insert ignore for MySQL https://github.com/lukas-krecan/ShedLock/commit/8a4ae7ad8103bb47f55d43bccf043ca261c24d7a
- Requires JDK 17
- Tested with Spring 6 (Spring Boot 3)
- Micronaut updated to 3.x.x
- R2DBC 1.x.x (still sucks)
- Spring Data 3.x.x
- Rudimentary support for CDI (tested with quarkus)
- New jOOQ lock provider
- SLF4j 2
- Deleted all deprecated code and support for old versions of libraries
- Better logging in JdbcTemplateProvider
- Dependency updates
- Deprecate old Couchbase lock provider
- Dependency updates
- Couchbase collection support (thanks @mesuutt)
- Dependency updates
- Fixed caching issues when the app is started by the DB does not exist yet (#1129)
- Dependency updates
- Introduced elasticsearch8 LockProvider and deperecated the orignal one (thanks @MarAra)
- Dependency updates
- ReactiveRedisLockProvider added (thanks @ericwcc)
- Dependency updates
- OpenSearch provider (thanks @Pinny3)
- Fix wrong reference to reactive Mongo in BOM #1048
- Dependency updates
- shedlock-bom module added
- Dependency updates
- Neo4j allows to specify database thanks @SergeyPlatonov
- Dependency updates
- Dropped support for Hazelcast <= 3 as it has unfixed vulnerability
- Dropped support for Spring Data Redis 1 as it is not supported
- Dependency updates
- memcached provider added (thanks @pinkhello)
- Dependency updates
- JDBC provider does not change autocommit attribute
- Dependency updates
- Jedis 4 lock provider
- Dependency updates
- In-memory lock provider added (thanks @kkocel)
- Dependency updates
- R2DBC support added (thanks @sokomishalov)
- Library upgrades
- Neo4j lock provider added (thanks @thimmwork)
- Library upgrades
- Ability to set transaction isolation in JdbcTemplateLockProvider
- Library upgrades
- KeepAliveLockProvider introduced
- Library upgrades
- LockExtender added
- Support for Apache Ignite (thanks @wirtsleg)
- Library upgrades
- Ability to set serialConsistencyLevel in Cassandra (thanks @DebajitKumarPhukan)
- Introduced shedlock-provider-jdbc-micronaut module (thanks @drmaas)
- Catching and logging Cassandra exception
- Support for custom keyspace in Cassandra provider
- Elastic unlock using IMMEDIATE refresh policy #422
- DB2 JDBC lock provider uses microseconds in DB time
- Various library upgrades
- Fixed DB JDBC server time #378
- Support for etcd (thanks grofoli)
- Fixed devtools compatibility #368
- Support for enhanced configuration in Cassandra provider (thanks DebajitKumarPhukan)
- LockConfigurationExtractor exposed as a Spring bean #359
- Handle CannotSerializeTransactionException #364
- Fixed Consul support for tokens and added enhanced Consul configuration (thanks DrWifey)
- Consul support for tokens
- Spring - EnableSchedulerLock.order param added to specify AOP proxy order
- JDBC - Log unexpected exceptions at ERROR level
- Hazelcast upgraded to 4.1
- Fix session leak in Consul provider #340 (thanks @haraldpusch)
- ArangoDB lock provider added (thanks @patrick-birkle)
- Support for Couchbase 3 driver (thanks @blitzenzzz)
- Removed forgotten configuration files form micronaut package (thanks @drmaas)
- Shutdown hook for Consul (thanks @kaliy)
- Support for Consul (thanks @kaliy)
- Various dependencies updated
- Deprecated default LockConfiguration constructor
- Lazy initialization of SqlStatementsSource #258
- MongoLockProvider uses mongodb-driver-sync
- Removed deprecated constructors from MongoLockProvider
- New Mongo reactive streams driver (thanks @codependent)
- Fixed JdbcTemplateLockProvider useDbTime() locking #244 thanks @gjorgievskivlatko
- Do not fail on DB type determining code if DB connection is not available
- Support for server time in DB2
- removed shedlock-provider-jdbc-internal module
- Support for server time in JdbcTemplateLockProvider
- Using custom non-null annotations
- Trimming time precision to milliseconds
- Micronaut upgraded to 1.3.4
- Add automatic DB tests for Oracle, MariaDB and MS SQL.
- DynamoDB 2 module introduced (thanks Mark Egan)
- JDBC template code refactored to not log error on failed insert in Postgres
- INSERT .. ON CONFLICT UPDATE is used for Postgres
- Make LockAssert.TestHelper public
- New module for Hazelcasts 4
- Ability to switch-off LockAssert in unit tests
- Support for Meta annotations and annotation inheritance in Spring
- Made compatible with PostgreSQL JDBC Driver 42.2.11
- Inject redis template
- ClockProvider introduced
- MongoLockProvider(MongoDatabase) introduced
- Support for non-void returning methods when PROXY_METHOD interception is used
- Introduced shedlock-provider-redis-spring-1 to make it work around Spring Data Redis 1 issue #105 (thanks @rygh4775)
- Jedis dependency upgraded to 3.2.0
- Support for JedisCluster
- Tests upgraded to JUnit 5
- Cassandra provider (thanks @mitjag)
- More configuration option for JdbcTemplateProvider
- Allow configuration of key prefix in RedisLockProvider #181 (thanks @krm1312)
- Fixed junit dependency scope #179
- Fix NPE caused by Redisson #178
- DefaultLockingTaskExecutor made reentrant #175
Version 4.0.0 is a major release changing quite a lot of stuff
net.javacrumbs.shedlock.core.SchedulerLock
has been replaced bynet.javacrumbs.shedlock.spring.annotation.SchedulerLock
. The original annotation has been in wrong module and was too complex. Please use the new annotation, the old one still works, but in few years it will be removed.- Default intercept mode changed from
PROXY_SCHEDULER
toPROXY_METHOD
. The reason is that there were a lot of issues withPROXY_SCHEDULER
(for example #168). You can still usePROXY_SCHEDULER
mode if you specify it manually. - Support for more readable duration strings
- Support for lock assertion
LockAssert.assertLocked()
- Support for Micronaut added
- Fixed bean definition configuration #171
EnableSchedulerLock.mode
renamed tointerceptMode
- Use standard Spring AOP configuration to honor Spring Boot config (supports
proxyTargetClass
flag) - Removed deprecated SpringLockableTaskSchedulerFactoryBean and related classes
- Removed support for XML configuration
- Updated dependency to Spring 2.1.9
- Support for lock extensions (beta)
- Zookeeper supports lockAtMostFor and lockAtLeastFor params
- Better debug logging
- Fixed potential deadlock in Hazelcast (thanks @HubertTatar)
- Finding class level annotation in proxy method mode (thanks @volkovs)
- ScheduledLockConfigurationBuilder deprecated
- LockProvides is initialized lazilly so it does not change DataSource initialization order
- MongoLockProvider accepts MongoCollection as a constructor param
- DynamoDBLockProvider added
- MongoLockProvider rewritten to use upsert
- ElasticsearchLockProvider added
- AOP proxy and annotation configuration support
- Can set Timezone to JdbcTemplateLock provider
- Support for Couchbase (thanks to @MoranVaisberg)
- Spring RedisLockProvider refactored to use RedisTemplate
- Support for transaction manager in JdbcTemplateLockProvider (thanks to @grmblfrz)
- Upgraded dependencies to Spring 5 and Spring Data 2
- Removed deprecated net.javacrumbs.shedlock.provider.jedis.JedisLockProvider (use net.javacrumbs.shedlock.provider.redis.jedis.JedisLockProvide instead)
- Removed deprecated SpringLockableTaskSchedulerFactory (use ScheduledLockConfigurationBuilder instead)
- ablility to clean lock cache
- shedlock-provider-redis-spring made compatible with spring-data-redis 1.x.x
- Added shedlock-provider-redis-spring (thanks to @siposr)
- shedlock-provider-jedis moved to shedlock-provider-redis-jedis
- Support for SPEL in lock name annotation
- Automatically closing TaskExecutor on Spring shutdown
- Removed spring-test from shedlock-spring compile time dependencies
- Added Automatic-Module-Names
- Hazelcast works with remote cluster
- Fixed ScheduledLockConfigurationBuilder interfaces #32
- Hazelcast code refactoring
- Support for Hazelcast (thanks to @peyo)
- Jedis constructor made more generic (thanks to @mgrzeszczak)
- Support for property placeholders in annotation lockAtMostForString/lockAtLeastForString
- Support for composed annotations
- ScheduledLockConfigurationBuilder introduced (deprecating SpringLockableTaskSchedulerFactory)
- Support for Redis (thanks to @clamey)
- Checking that lockAtMostFor is in the future
- Checking that lockAtMostFor is larger than lockAtLeastFor
- jdbc-template-provider does not participate in task transaction
- Support for @SchedulerLock annotations on proxied classes
- LockableTaskScheduler made AutoClosable so it's closed upon Spring shutdown
- Support for lockAtLeastFor
- Possible to configure defaultLockFor time so it does not have to be repeated in every annotation
- ZooKeeper nodes created under /shedlock by default
- JdbcLockProvider insert does not fail on DataIntegrityViolationException
- Extracted LockingTaskExecutor
- LockManager.executeIfNotLocked renamed to executeWithLock
- Default table name in JDBC lock providers
@ShedlulerLock.name
made obligatory@ShedlulerLock.lockForMillis
renamed to lockAtMostFor- Adding plain JDBC LockProvider
- Adding ZooKeepr LockProvider