Configuring Infinispan caches

The cache mode that you should choose depends on the qualities and guarantees you need for your data.

The following table summarizes the primary differences between cache modes:

A simple cache is a type of local cache that disables support for the following capabilities:

However, you can use other Infinispan capabilities with simple caches such as expiration, eviction, statistics, and security features. If you configure a capability that is not compatible with a simple cache, Infinispan throws an exception.

Even with synchronous replication, distributed caches are not linearizable. For transactional caches, they do not support serialization/snapshot isolation.

For example, a thread is carrying out a single put request:

But another thread might see the values in a different order:

The reason is that read can return the value from any owner, depending on how fast the primary owner replies. The write is not atomic across all the owners. In fact, the primary commits the update only after it receives a confirmation from the backup. While the primary is waiting for the confirmation message from the backup, reads from the backup will see the new value, but reads from the primary will see the old one.

Distributed caches split entries into a fixed number of segments and assign each segment to a list of owner nodes. Replicated caches do the same, with the exception that every node is an owner.

The first node in the list of owners is the primary owner. The other nodes in the list are backup owners. When the cache topology changes, because a node joins or leaves the cluster, the segment ownership table is broadcast to every node. This allows nodes to locate keys without making multicast requests or maintaining metadata for each key.

The numSegments property configures the number of segments available. However, the number of segments cannot change unless the cluster is restarted.

Likewise the key-to-segment mapping cannot change. Keys must always map to the same segments regardless of cluster topology changes. It is important that the key-to-segment mapping evenly distributes the number of segments allocated to each node while minimizing the number of segments that must move when the cluster topology changes.

You can configure ConsistentHashFactory implementations, including custom ones, with embedded caches only.

Capacity factors allocate the number of segments based on resources available to each node in the cluster.

The capacity factor for a node applies to segments for which that node is both the primary owner and backup owner. In other words, the capacity factor specifies is the total capacity that a node has in comparison to other nodes in the cluster.

The default value is 1 which means that all nodes in the cluster have an equal capacity and Infinispan allocates the same number of segments to all nodes in the cluster.

However, if nodes have different amounts of memory available to them, you can configure the capacity factor so that the Infinispan hashing algorithm assigns each node a number of segments weighted by its capacity.

The value for the capacity factor configuration must be a positive number and can be a fraction such as 1.5. You can also configure a capacity factor of 0 but is recommended only for nodes that join the cluster temporarily and should use the zero capacity configuration instead.

Infinispan nodes create local replicas when they retrieve entries from another node in the cluster. L1 caches avoid repeatedly looking up entries on primary owner nodes and adds performance.

The following diagram illustrates how L1 caches work:

In the "L1 cache" diagram:

Enabling L1 improves performance for read operations but requires primary owner nodes to broadcast invalidation messages when entries are modified. This ensures that Infinispan removes any out of date replicas across the cluster. However this also decreases performance of write operations and increases memory usage, reducing overall capacity of caches.

Server hinting increases availability of data in distributed caches by replicating entries across as many servers, racks, and data centers as possible.

When Infinispan distributes the copies of your data, it follows the order of precedence: site, rack, machine, and node. All of the configuration attributes are optional. For example, when you specify only the rack IDs, then Infinispan distributes the copies across different racks and nodes.

Server hinting can impact cluster rebalancing operations by moving more segments than necessary if the number of segments for the cache is too low.

In a distributed cache, a key is allocated to a list of nodes with an opaque algorithm. There is no easy way to reverse the computation and generate a key that maps to a particular node. However, Infinispan can generate a sequence of (pseudo-)random keys, see what their primary owner is, and hand them out to the application when it needs a key mapping to a particular node.

Following code snippet depicts how a reference to this service can be obtained and used.

The service is started at step 2: after this point it uses the supplied Executor to generate and queue keys. At step 3, we obtain a key from the service, and at step 4 we use it.

KeyAffinityService extends Lifecycle, which allows stopping and (re)starting it:

The service is instantiated through KeyAffinityServiceFactory. All the factory methods have an Executor parameter, that is used for asynchronous key generation (so that it won’t happen in the caller’s thread). It is the user’s responsibility to handle the shutdown of this Executor.

The KeyAffinityService, once started, needs to be explicitly stopped. This stops the background key generation and releases other held resources.

The only situation in which KeyAffinityService stops by itself is when the Cache Manager with which it was registered is shutdown.

When the cache topology changes, the ownership of the keys generated by the KeyAffinityService might change. The key affinity service keep tracks of these topology changes and doesn’t return keys that would currently map to a different node, but it won’t do anything about keys generated earlier.

As such, applications should treat KeyAffinityService purely as an optimization, and they should not rely on the location of a generated key for correctness.

In particular, applications should not rely on keys generated by KeyAffinityService for the same address to always be located together. Collocation of keys is only provided by the Grouping API.

Complementary to the Key affinity service, the Grouping API allows you to co-locate a group of entries on the same nodes, but without being able to select the actual nodes.

By default, the segment of a key is computed using the key’s hashCode(). If you use the Grouping API, Infinispan will compute the segment of the group and use that as the segment of the key.

When the Grouping API is in use, it is important that every node can still compute the owners of every key without contacting other nodes. For this reason, the group cannot be specified manually. The group can either be intrinsic to the entry (generated by the key class) or extrinsic (generated by an external function).

To use the Grouping API, you must enable groups.

If you have control of the key class (you can alter the class definition, it’s not part of an unmodifiable library), then we recommend using an intrinsic group. The intrinsic group is specified by adding the @Group annotation to a method, for example:

If you don’t have control over the key class, or the determination of the group is an orthogonal concern to the key class, we recommend using an extrinsic group. An extrinsic group is specified by implementing the Grouper interface.

If multiple Grouper classes are configured for the same key type, all of them will be called, receiving the value computed by the previous one. If the key class also has a @Group annotation, the first Grouper will receive the group computed by the annotated method. This allows you even greater control over the group when using an intrinsic group.

Grouper implementations must be registered explicitly in the cache configuration. If you are configuring Infinispan programmatically:

Or, if you are using XML:

AdvancedCache has two group-specific methods:

Both methods iterate over the entire data container and store (if present), so they can be slow when a cache contains lots of small groups.

Because the Cache interface extends java.util.Map, write methods like put(key, value) and remove(key) return the previous value by default.

In some cases, the return value may not be correct:

Transactional caches return the correct previous value in cases 3 and 4. However, transactional caches also have a gotcha: in distributed mode, the read-committed isolation level is implemented as repeatable-read. That means this example of "double-checked locking" won’t work:

The correct way to implement this is to use cache.getAdvancedCache().withFlags(Flag.FORCE_WRITE_LOCK).get(k).

In caches with optimistic locking, writes can also return stale previous values. Write skew checks can avoid stale previous values.

You can create declarative cache configuration in XML, JSON, and YAML format.

All declarative caches must conform to the Infinispan schema. Configuration in JSON format must follow the structure of an XML configuration, elements correspond to objects and attributes correspond to fields.

Create caches from configuration templates.

Cache configuration templates can inherit from other templates to extend and override settings.

Cache template inheritance is hierarchical. For a child configuration template to inherit from a parent, you must include it after the parent template.

Additionally, template inheritance is additive for elements that have multiple values. A cache that inherits from another template merges the values from that template, which can override properties.

You can add wildcards to cache configuration template names. If you then create caches where the name matches the wildcard, Infinispan applies the configuration template.

Using the preceding example, if you create a cache named "async-dist-cache-prod" then Infinispan uses the configuration from the async-dist-cache-* template.

Split cache configuration templates into multiple XML files for granular flexibility and reference them with XML inclusions (XInclude).

Infinispan also provides an infinispan-config-fragment-15.0.xsd schema that you can use with configuration fragments.

Infinispan Server provides a default Cache Manager that controls the lifecycle of remote caches. Starting Infinispan Server automatically instantiates the Cache Manager so you can create and delete remote caches and other resources like Protobuf schema.

After you start Infinispan Server and add user credentials, you can view details about the Cache Manager and get cluster information from Infinispan Console.

You can also get information about the Cache Manager through the Command Line Interface (CLI) or REST API:

Use Infinispan Console to create remote caches in an intuitive visual interface from any web browser.

Use the Infinispan Command Line Interface (CLI) to add remote caches on Infinispan Server.

Use the Infinispan Hot Rod API to create remote caches on Infinispan Server from Java, C++, .NET/C#, JS clients and more.

This procedure shows you how to use Hot Rod Java clients that create remote caches on first access. You can find code examples for other Hot Rod clients in the Infinispan Tutorials.

Use the Infinispan REST API to create remote caches on Infinispan Server from any suitable HTTP client.

Add Infinispan to your project to create embedded caches in your applications.

Infinispan provides a GlobalConfigurationBuilder API that controls the Cache Manager and a ConfigurationBuilder API that configures caches.

Invoke the getCache(String) method to obtain caches, as follows:

The preceding operation creates a cache named myCache, if it does not already exist, and returns it.

Using the getCache() method creates the cache only on the node where you invoke the method. In other words, it performs a local operation that must be invoked on each node across the cluster. Typically, applications deployed across multiple nodes obtain caches during initialization to ensure that caches are symmetric and exist on each node.

Invoke the createCache() method to create caches dynamically across the entire cluster.

The preceding operation also automatically creates caches on any nodes that subsequently join the cluster.

Caches that you create with the createCache() method are ephemeral by default. If the entire cluster shuts down, the cache is not automatically created again when it restarts.

Use the PERMANENT flag to ensure that caches can survive restarts.

For the PERMANENT flag to take effect, you must enable global state and set a configuration storage provider.

For more information about configuration storage providers, see GlobalStateConfigurationBuilder#configurationStorage().

Infinispan provides a Cache interface that exposes simple methods for adding, retrieving and removing entries, including atomic mechanisms exposed by the JDK’s ConcurrentMap interface. Based on the cache mode used, invoking these methods will trigger a number of things to happen, potentially even including replicating an entry to a remote node or looking up an entry from a remote node, or potentially a cache store.

For simple usage, using the Cache API should be no different from using the JDK Map API, and hence migrating from simple in-memory caches based on a Map to Infinispan’s Cache should be trivial.

Certain methods exposed in Map have certain performance consequences when used with Infinispan, such as size() , values() , keySet() and entrySet() . Specific methods on the keySet, values and entrySet are fine for use please see their Javadoc for further details.

Attempting to perform these operations globally would have large performance impact as well as become a scalability bottleneck. As such, these methods should only be used for informational or debugging purposes only.

It should be noted that using certain flags with the withFlags() method can mitigate some of these concerns, please check each method’s documentation for more details.

Further to simply storing entries, Infinispan’s cache API allows you to attach mortality information to data. For example, simply using put(key, value) would create an immortal entry, i.e., an entry that lives in the cache forever, until it is removed (or evicted from memory to prevent running out of memory). If, however, you put data in the cache using put(key, value, lifespan, timeunit) , this creates a mortal entry, i.e., an entry that has a fixed lifespan and expires after that lifespan.

In addition to lifespan , Infinispan also supports maxIdle as an additional metric with which to determine expiration. Any combination of lifespans or maxIdles can be used.

Infinispan’s Cache class contains a different 'put' operation called putForExternalRead . This operation is particularly useful when Infinispan is used as a temporary cache for data that is persisted elsewhere. Under heavy read scenarios, contention in the cache should not delay the real transactions at hand, since caching should just be an optimization and not something that gets in the way.

To achieve this, putForExternalRead() acts as a put call that only operates if the key is not present in the cache, and fails fast and silently if another thread is trying to store the same key at the same time. In this particular scenario, caching data is a way to optimise the system and it’s not desirable that a failure in caching affects the on-going transaction, hence why failure is handled differently. putForExternalRead() is considered to be a fast operation because regardless of whether it’s successful or not, it doesn’t wait for any locks, and so returns to the caller promptly.

To understand how to use this operation, let’s look at basic example. Imagine a cache of Person instances, each keyed by a PersonId , whose data originates in a separate data store. The following code shows the most common pattern of using putForExternalRead within the context of this example:

Note that putForExternalRead should never be used as a mechanism to update the cache with a new Person instance originating from application execution (i.e. from a transaction that modifies a Person’s address). When updating cached values, please use the standard put operation, otherwise the possibility of caching corrupt data is likely.

Provide unique remote JMX ports to expose Infinispan MBeans through connections in JMXServiceURL format.

You can enable remote JMX ports using one of the following approaches:

Start Infinispan Server with a remote JMX port enabled using one of the following ways:

Infinispan exposes JMX MBeans that represent manageable resources.

For a complete list of available JMX MBeans along with descriptions and available operations and attributes, see the Infinispan JMX Components documentation.

Infinispan includes an MBeanServerLookup interface that you can use to register MBeans in custom MBeanServer instances.

When you configure Infinispan eviction you specify:

You can either perform eviction manually or configure Infinispan to do one of the following:

Refer to the schema reference for more details about the eviction strategies.

Limit the size of Infinispan caches to a total number of entries.

In the following example, Infinispan removes an entry when the cache contains a total of 500 entries and a new entry is created:

Limit the size of Infinispan caches to a maximum amount of memory.

In the following example, Infinispan removes an entry when the size of the cache reaches 1.5 GB (gigabytes) and a new entry is created:

If you choose the manual eviction strategy, Infinispan does not perform eviction. You must do so manually with the evict() method.

You should use manual eviction with embedded caches only. For remote caches, you should always configure Infinispan with the REMOVE or EXCEPTION eviction strategy.

Passivation persists data to cache stores when Infinispan evicts entries. You should always enable eviction if you enable passivation, as in the following examples:

When you configure expiration, Infinispan stores keys with metadata that determines when entries expire.

Infinispan checks if lifespan or maximum idle metadata is set and then compares the values with the current time.

If (creation + lifespan < currentTime) or (lastUsed + maxIdle < currentTime) then Infinispan detects that the entry is expired.

Expiration occurs whenever entries are accessed or found by the expiration reaper.

For example, k1 reaches the maximum idle time and a client makes a Cache.get(k1) request. In this case, Infinispan detects that the entry is expired and removes it from the data container. The Cache.get(k1) request returns null.

Infinispan also expires entries from cache stores, but only with lifespan expiration. Maximum idle expiration does not work with cache stores. In the case of cache loaders, Infinispan cannot expire entries because loaders can only read from external storage.

Infinispan uses a reaper thread that runs periodically to detect and remove expired entries. The expiration reaper ensures that expired entries that are no longer accessed are removed.

The Infinispan ExpirationManager interface handles the expiration reaper and exposes the processExpiration() method.

In some cases, you can disable the expiration reaper and manually expire entries by calling processExpiration(); for instance, if you are using local cache mode with a custom application where a maintenance thread runs periodically.

Because maximum idle expiration relies on the last access time for cache entries, it has some limitations with clustered cache modes.

With lifespan expiration, the creation time for cache entries provides a value that is consistent across clustered caches. For example, the creation time for k1 is always the same on all nodes.

For maximum idle expiration with clustered caches, last access time for entries is not always the same on all nodes. To ensure that entries have the same relative access times across clusters, Infinispan sends touch commands to all owners when keys are accessed.

The touch commands that Infinispan send have the following considerations:

Set lifespan and maximum idle times for all entries in a cache.

In the following example, Infinispan expires all cache entries after 5 seconds or 1 second after the last access time, whichever happens first:

Specify lifespan and maximum idle times for individual entries. When you add lifespan and maximum idle times to entries, those values take priority over expiration configuration for caches.

When you add entries to off-heap caches, Infinispan dynamically allocates native memory to your data.

Infinispan hashes the serialized byte[] for each key into buckets that are similar to a standard Java HashMap. Buckets include address pointers that Infinispan uses to locate entries that you store in off-heap memory.

Infinispan determines equality of Java objects in off-heap storage using the serialized byte[] representation of each object instead of the object instance.

Infinispan uses an array of locks to protect off-heap address spaces. The number of locks is twice the number of cores and then rounded to the nearest power of two. This ensures that there is an even distribution of ReadWriteLock instances to prevent write operations from blocking read operations.

Configure Infinispan to store cache entries in native memory outside the JVM heap space.

Infinispan stores cache entries as bytes in native memory. Eviction happens when there are 100 entries in the data container and Infinispan gets a request to create a new entry:

Writes to data in memory result in writes to persistent storage.

If Infinispan evicts data from memory, then data in persistent storage includes entries that are evicted from memory. In this way persistent storage is a superset of the in-memory cache. This is recommended when you require highest consistency as the store will be able to be read again after a crash.

If you do not configure eviction, then data in persistent storage provides a copy of data in memory.

Infinispan adds data to persistent storage only when it evicts data from memory, an entry is removed or upon shutting down the node.

When Infinispan activates entries, it restores data in memory but keeps the data in the store still. This allows for writes to be just as fast as without a store, and still maintains consistency. When an entry is created or updated only the in memory will be updated and thus the store will be outdated for the time being.

To gurantee data consistency any store that is not shared should always have purgeOnStartup enabled. This is true for both passivation enabled or disabled since a store could hold an outdated entry while down and resurrect it at a later point.

The following table shows data in memory and in persistent storage after a series of operations:

Enable and configure the location where Infinispan stores global state for clustered embedded caches.

Add file-based cache stores to Infinispan to persist data on the host filesystem.

If required, you can configure Infinispan to create single file stores.

Create managed datasources as part of your Infinispan Server configuration to optimize connection pooling and performance for JDBC database connections. You can then specify the JDNI name of the managed datasources in your caches, which centralizes JDBC connection configuration for your deployment.

Use the Infinispan Command Line Interface (CLI) to test the datasource connection, as follows:

You can use a properties file to configure pooled connection factories for JDBC string-based cache stores.

Infinispan loads keys and values from columns in database tables via SQL cache stores, automatically using the appropriate data types. The following CREATE statement adds a table named "books" that has two columns, isbn and title:

When you use this table with a SQL cache store, Infinispan adds an entry to the cache using the isbn column as the key and the title column as the value.

Add a SQL table cache store to your configuration if you want Infinispan to load data from a database table. When it connects to the database, Infinispan uses metadata from the table to detect column names and data types. Infinispan also automatically determines which columns in the database are part of the primary key.

The following example loads a distributed cache from a database table named "books" using composite values defined in a Protobuf schema:

SQL query cache stores let you load caches from multiple database tables, including from sub-columns in database tables, and perform insert, update, and delete operations.

Find out about common issues and errors with SQL cache stores and how to troubleshoot them.

Infinispan logs this message in the following cases:

To resolve this issue you should:

Configure Infinispan caches with JDBC string-based cache stores that can connect to databases.

The Infinispan Service Provider Interface (SPI) enables read and write operations to external storage through the NonBlockingStore interface and has the following features:

Create custom cache stores with implementations of the NonBlockingStore API.

The following are examples show how to configure Infinispan with custom cache store implementations:

To use your cache store implementation with Infinispan Server, you must provide it with a JAR file.

Infinispan provides the CLI migrate store command that recreates data for the latest Infinispan cache store implementations.

The store migrator takes a cache store from a previous version of Infinispan as source and uses a cache store implementation as target.

When you run the store migrator, it creates the target cache with the cache store type that you define using the EmbeddedCacheManager interface. The store migrator then loads entries from the source store into memory and then puts them into the target cache.

The store migrator also lets you migrate data from one type of cache store to another. For example, you can migrate from a JDBC string-based cache store to a SIFS cache store.

Use the migrator.properties file to configure properties for source and target cache stores.

Run the store migrator to migrate data from one cache store to another.

Network outages or errors that cause Infinispan clusters to split into partitions can result in data loss or consistency issues regardless of any handling strategy or merge policy.

If a write operation takes place on a node that is in a minor partition when a split occurs, and before Infinispan detects the split, that value is lost when Infinispan transfers state to that minor partition during the merge.

In the event that all partitions are in the DEGRADED mode that value is not lost because no state transfer occurs but the entry can have an inconsistent value. For transactional caches write operations that are in progress when the split occurs can be committed on some nodes and rolled back on other nodes, which also results in inconsistent values.

During the split and the time that Infinispan detects it, it is possible to get stale reads from a cache in a minor partition that has not yet entered DEGRADED mode.

When Infinispan starts removing partitions nodes reconnect to the cluster with a series of merge events. Before this merge process completes it is possible that write operations on transactional caches succeed on some nodes but not others, which can potentially result in stale reads until the entries are updated.

This topic illustrates how Infinispan recovers from split clusters with caches that use the DENY_READ_WRITES partition handling strategy.

As an example, a Infinispan cluster has four nodes and includes a distributed cache with two replicas for each entry (owners=2). There are four entries in the cache, k1, k2, k3 and k4.

With the DENY_READ_WRITES strategy, if the cluster splits into partitions, Infinispan allows cache operations only if all replicas of an entry are in the same partition.

In the following diagram, while the cache is split into partitions, Infinispan allows read and write operations for k1 on partition 1 and k4 on partition 2. Because there is only one replica for k2 and k3 on either partition 1 or partition 2, Infinispan denies read and write operations for those entries.

When network conditions allow the nodes to re-join the same cluster view, Infinispan merges the partitions without state transfer and restores normal cache operations.

Determine if caches on Infinispan clusters are in AVAILABLE mode or DEGRADED mode during a network partition.

When Infinispan clusters split into partitions, nodes in those partitions can enter DEGRADED mode to guarantee data consistency. In DEGRADED mode clusters do not allow cache operations resulting in loss of availability.

Verify availability of clustered caches in network partitions in one of the following ways:

Make caches available for read and write operations by forcing them out of DEGRADED mode.

Make caches available in one of the following ways:

User roles are sets of permissions with different access levels.

Infinispan implements users as a collection of principals. Principals represent either an individual user identity, such as a username, or a group to which the users belong. Internally, these are implemented with the javax.security.auth.Subject class.

To enable authorization, the principals must be mapped to role names, which are then expanded into a set of permissions.

Infinispan includes the PrincipalRoleMapper API for associating security principals to roles, and the RolePermissionMapper API for associating roles with specific permissions.

Infinispan provides the following role and permission mapper implementations:

Infinispan enables the cluster role mapper and cluster permission mapper by default. To use a different implementation for role mapping, you must configure the role mappers.

With embedded caches you can programmatically configure role and permission mappers with the principalRoleMapper() and rolePermissionMapper() methods.

The cluster role mapper maintains a dynamic mapping between principals and roles. The cluster permission mapper maintains a dynamic set of role definitions. In both cases, the mappings are stored in the cluster registry and can be manipulated at runtime using either the CLI or the REST API.

Transactions are configured at cache level. Below is the configuration that affects a transaction behaviour and a small description of each configuration attribute.

or programmatically:

For more details on how Two-Phase-Commit (2PC) is implemented in Infinispan and how locks are being acquired see the section below. More details about the configuration settings are available in Configuration reference.

Infinispan offers two isolation levels - READ_COMMITTED and REPEATABLE_READ.

These isolation levels determine when readers see a concurrent write, and are internally implemented using different subclasses of MVCCEntry, which have different behaviour in how state is committed back to the data container.

Here’s a more detailed example that should help understand the difference between READ_COMMITTED and REPEATABLE_READ in the context of Infinispan. With READ_COMMITTED, if between two consecutive read calls on the same key, the key has been updated by another transaction, the second read may return the new updated value:

With REPEATABLE_READ, the final get will still return v. So, if you’re going to retrieve the same key multiple times within a transaction, you should use REPEATABLE_READ.

However, as read-locks are not acquired even for REPEATABLE_READ, this phenomena can occur:

Write skews occur when two transactions independently and simultaneously read and write to the same key. The result of a write skew is that both transactions successfully commit updates to the same key but with different values.

Infinispan automatically performs write skew checks to ensure data consistency for REPEATABLE_READ isolation levels in optimistic transactions. This allows Infinispan to detect and roll back one of the transactions.

When operating in LOCAL mode, write skew checks rely on Java object references to compare differences, which provides a reliable technique for checking for write skews.

If a CacheException (or a subclass of it) is thrown by a cache method within the scope of a JTA transaction, then the transaction is automatically marked for rollback.

By default Infinispan registers itself as a first class participant in distributed transactions through XAResource. There are situations where Infinispan is not required to be a participant in the transaction, but only to be notified by its lifecycle (prepare, complete): e.g. in the case Infinispan is used as a 2nd level cache in Hibernate.

Infinispan allows transaction enlistment through Synchronization. To enable it just use NON_XA transaction mode.

Synchronizations have the advantage that they allow TransactionManager to optimize 2PC with a 1PC where only one other resource is enlisted with that transaction (last resource commit optimization). E.g. Hibernate second level cache: if Infinispan registers itself with the TransactionManager as a XAResource than at commit time, the TransactionManager sees two XAResource (cache and database) and does not make this optimization. Having to coordinate between two resources it needs to write the tx log to disk. On the other hand, registering Infinispan as a Synchronization makes the TransactionManager skip writing the log to the disk (performance improvement).

Batching allows atomicity and some characteristics of a transaction, but not full-blown JTA or XA capabilities. Batching is often a lot lighter and cheaper than a full-blown transaction.

Recovery is a feature of XA transactions, which deal with the eventuality of a resource or possibly even the transaction manager failing, and recovering accordingly from such a situation.

In Infinispan clusters, primary owner nodes are responsible for locking keys.

For non-transactional caches, Infinispan forwards the write operation to the primary owner of the key so it can attempt to lock it. Infinispan either then forwards the write operation to the other owners or throws an exception if it cannot lock the key.

For transactional caches, primary owners can lock keys with optimistic and pessimistic locking modes. Infinispan also supports different isolation levels to control concurrent reads between transactions.

The LockManager is a component that is responsible for locking an entry for writing. The LockManager makes use of a LockContainer to locate/hold/create locks. LockContainers come in two broad flavours, with support for lock striping and with support for one lock per entry.

Lock striping entails the use of a fixed-size, shared collection of locks for the entire cache, with locks being allocated to entries based on the entry’s key’s hash code. Similar to the way the JDK’s ConcurrentHashMap allocates locks, this allows for a highly scalable, fixed-overhead locking mechanism in exchange for potentially unrelated entries being blocked by the same lock.

The alternative is to disable lock striping - which would mean a new lock is created per entry. This approach may give you greater concurrent throughput, but it will be at the cost of additional memory usage, garbage collection churn, etc.

The size of the shared lock collection used by lock striping can be tuned using the concurrencyLevel attribute of the <locking /> configuration element.

Configuration example:

Or

In addition to determining the size of the striped lock container, this concurrency level is also used to tune any JDK ConcurrentHashMap based collections where related, such as internal to DataContainers. Please refer to the JDK ConcurrentHashMap Javadocs for a detailed discussion of concurrency levels, as this parameter is used in exactly the same way in Infinispan.

Configuration example:

Or

The lock timeout specifies the amount of time, in milliseconds, to wait for a contented lock.

Configuration example:

Or

The fact that a single owner is locked (as opposed to all owners being locked) does not break the following consistency guarantee: if key K is hashed to nodes {A, B} and transaction TX1 acquires a lock for K, let’s say on A. If another transaction, TX2, is started on B (or any other node) and TX2 tries to lock K then it will fail with a timeout as the lock is already held by TX1. The reason for this is the that the lock for a key K is always, deterministically, acquired on the same node of the cluster, regardless of where the transaction originates.

Infinispan supports two forms of data versioning: simple and external. The simple versioning is used in transactional caches for write skew check.

The external versioning is used to encapsulate an external source of data versioning within Infinispan, such as when using Infinispan with Hibernate which in turn gets its data version information directly from a database.

In this scheme, a mechanism to pass in the version becomes necessary, and overloaded versions of put() and putForExternalRead() will be provided in AdvancedCache to take in an external data version. This is then stored on the InvocationContext and applied to the entry at commit time.