/* * Copyright (c) 2014 Cisco Systems, Inc. and others. All rights reserved. * * This program and the accompanying materials are made available under the * terms of the Eclipse Public License v1.0 which accompanies this distribution, * and is available at http://www.eclipse.org/legal/epl-v10.html */ package org.opendaylight.controller.md.sal.common.api.data; import org.opendaylight.controller.md.sal.common.api.TransactionStatus; import org.opendaylight.yangtools.concepts.Path; import org.opendaylight.yangtools.yang.common.RpcResult; import com.google.common.util.concurrent.ListenableFuture; /** * Write transaction provides mutation capabilities for a data tree. * *

* Initial state of write transaction is a stable snapshot of the current data tree. * The state is captured when the transaction is created and its state and underlying * data tree are not affected by other concurrently running transactions. *

* Write transactions are isolated from other concurrent write transactions. All * writes are local to the transaction and represent only a proposal of state * change for the data tree and it is not visible to any other concurrently running * transaction. *

* Applications publish the changes proposed in the transaction by calling {@link #commit} * on the transaction. This seals the transaction * (preventing any further writes using this transaction) and submits it to be * processed and applied to global conceptual data tree. *

* The transaction commit may fail due to a concurrent transaction modifying and committing data in * an incompatible way. See {@link #commit()} for more concrete commit failure examples. * * *

* Implementation Note: This interface is not intended to be implemented * by users of MD-SAL, but only to be consumed by them. * * @param

* Type of path (subtree identifier), which represents location in * tree * @param * Type of data (payload), which represents data payload */ public interface AsyncWriteTransaction

, D> extends AsyncTransaction { /** * Cancels the transaction. * * Transactions can only be cancelled if it's status is * {@link TransactionStatus#NEW} or {@link TransactionStatus#SUBMITED} * * Invoking cancel() on {@link TransactionStatus#FAILED} or * {@link TransactionStatus#CANCELED} will have no effect. * * @throws IllegalStateException * If transaction status is {@link TransactionStatus#COMMITED} * */ public void cancel(); /** * Store a piece of data at specified path. This acts as an add / replace * operation, which is to say that whole subtree will be replaced by * specified path. Performing the following put operations: * *

     * 1) container { list [ a ] }
     * 2) container { list [ b ] }
     * 
* * will result in the following data being present: * *
     * container { list [ b ] }
     * 
* * * If you need to make sure that a parent object exists, but you do not want modify * its preexisting state by using put, consider using * {@link #merge(LogicalDatastoreType, Path, Object)} * * @param store * Logical data store which should be modified * @param path * Data object path * @param data * Data object to be written to specified path * @throws IllegalStateException * if the transaction is no longer {@link TransactionStatus#NEW} */ public void put(LogicalDatastoreType store, P path, D data); /** * Store a piece of data at the specified path. This acts as a merge operation, * which is to say that any pre-existing data which is not explicitly * overwritten will be preserved. This means that if you store a container, * its child lists will be merged. Performing the following merge * operations: * *
     * 1) container { list [ a ] }
     * 2) container { list [ b ] }
     * 
* * will result in the following data being present: * *
     * container { list [ a, b ] }
     * 
* * This also means that storing the container will preserve any * augmentations which have been attached to it. *

* If you require an explicit replace operation, use * {@link #put(LogicalDatastoreType, Path, Object)} instead. * * @param store * Logical data store which should be modified * @param path * Data object path * @param data * Data object to be written to specified path * @throws IllegalStateException * if the transaction is no longer {@link TransactionStatus#NEW} */ public void merge(LogicalDatastoreType store, P path, D data); /** * Remove a piece of data from specified path. This operation does not fail * if the specified path does not exist. * * @param store * Logical data store which should be modified * @param path * Data object path * @throws IllegalStateException * if the transaction is no longer {@link TransactionStatus#NEW} */ public void delete(LogicalDatastoreType store, P path); /** * * Closes transaction and resources allocated to the transaction. * * This call does not change Transaction status. Client SHOULD explicitly * {@link #commit()} or {@link #cancel()} transaction. * * @throws IllegalStateException * if the transaction has not been updated by invoking * {@link #commit()} or {@link #cancel()}. */ @Override public void close(); /** * Submits transaction to be applied to update logical data tree. *

* This call logically seals the transaction, which prevents the client from * further changing data tree using this transaction. Any subsequent calls to * {@link #put(LogicalDatastoreType, Path, Object)}, * {@link #merge(LogicalDatastoreType, Path, Object)} or * {@link #delete(LogicalDatastoreType, Path)} will fail with * {@link IllegalStateException}. * * The transaction is marked as {@link TransactionStatus#SUBMITED} and * enqueued into the data store backed for processing. * *

* Whether or not the commit is successful is determined by versioning * of data tree and validation of registered commit participants * {@link AsyncConfigurationCommitHandler} * if transaction changes {@link LogicalDatastoreType#CONFIGURATION} data tree. *

* The effects of successful commit of data depends on * other data change listeners {@link AsyncDataChangeListener} and * {@link AsyncConfigurationCommitHandler}, which was registered to the * same {@link AsyncDataBroker}, to which this transaction belongs. * *

Failure scenarios

*

* Transaction may fail because of multiple reasons, such as *

* *

Change compatibility

* * There are several sets of changes which could be considered incompatible * between two transactions which are derived from same initial state. * Rules for conflict detection applies recursively for each subtree * level. * *

Change compatibility of leafs, leaf-list items

* * Following table shows state changes and failures between two concurrent transactions, * which are based on same initial state, Tx 1 completes successfully * before Tx 2 is submitted. * * * * * * * * * * * * * * * * * *
Initial stateTx 1Tx 2Result
Emptyput(A,1)put(A,2)Tx 2 will fail, state is A=1
Emptyput(A,1)merge(A,2)A=2
Emptymerge(A,1)put(A,2)Tx 2 will fail, state is A=1
Emptymerge(A,1)merge(A,2)A=2
A=0put(A,1)put(A,2)Tx 2 will fail, A=1
A=0put(A,1)merge(A,2)A=2
A=0merge(A,1)put(A,2)Tx 2 will fail, A=1
A=0merge(A,1)merge(A,2)A=2
A=0delete(A)put(A,2)Tx 2 will fail, A does not exists
A=0delete(A)merge(A,2)A=2
* *

Change compatibility of subtrees

* * Following table shows state changes and failures between two concurrent transactions, * which are based on same initial state, Tx 1 completes successfully * before Tx 2 is submitted. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Initial stateTx 1Tx 2Result
Emptyput(TOP,[])put(TOP,[])Tx 2 will fail, state is TOP=[]
Emptyput(TOP,[])merge(TOP,[])TOP=[]
Emptyput(TOP,[FOO=1])put(TOP,[BAR=1])Tx 2 will fail, state is TOP=[FOO=1]
Emptyput(TOP,[FOO=1])merge(TOP,[BAR=1])TOP=[FOO=1,BAR=1]
Emptymerge(TOP,[FOO=1])put(TOP,[BAR=1])Tx 2 will fail, state is TOP=[FOO=1]
Emptymerge(TOP,[FOO=1])merge(TOP,[BAR=1])TOP=[FOO=1,BAR=1]
TOP=[]put(TOP,[FOO=1])put(TOP,[BAR=1])Tx 2 will fail, state is TOP=[FOO=1]
TOP=[]put(TOP,[FOO=1])merge(TOP,[BAR=1])state is TOP=[FOO=1,BAR=1]
TOP=[]merge(TOP,[FOO=1])put(TOP,[BAR=1])Tx 2 will fail, state is TOP=[FOO=1]
TOP=[]merge(TOP,[FOO=1])merge(TOP,[BAR=1])state is TOP=[FOO=1,BAR=1]
TOP=[]delete(TOP)put(TOP,[BAR=1])Tx 2 will fail, state is empty store
TOP=[]delete(TOP)merge(TOP,[BAR=1])state is TOP=[BAR=1]
TOP=[]put(TOP/FOO,1)put(TOP/BAR,1])state is TOP=[FOO=1,BAR=1]
TOP=[]put(TOP/FOO,1)merge(TOP/BAR,1)state is TOP=[FOO=1,BAR=1]
TOP=[]merge(TOP/FOO,1)put(TOP/BAR,1)state is TOP=[FOO=1,BAR=1]
TOP=[]merge(TOP/FOO,1)merge(TOP/BAR,1)state is TOP=[FOO=1,BAR=1]
TOP=[]delete(TOP)put(TOP/BAR,1)Tx 2 will fail, state is empty store
TOP=[]delete(TOP)merge(TOP/BAR,1]Tx 2 will fail, state is empty store
TOP=[FOO=1]put(TOP/FOO,2)put(TOP/BAR,1)state is TOP=[FOO=2,BAR=1]
TOP=[FOO=1]put(TOP/FOO,2)merge(TOP/BAR,1)state is TOP=[FOO=2,BAR=1]
TOP=[FOO=1]merge(TOP/FOO,2)put(TOP/BAR,1)state is TOP=[FOO=2,BAR=1]
TOP=[FOO=1]merge(TOP/FOO,2)merge(TOP/BAR,1)state is TOP=[FOO=2,BAR=1]
TOP=[FOO=1]delete(TOP/FOO)put(TOP/BAR,1)state is TOP=[BAR=1]
TOP=[FOO=1]delete(TOP/FOO)merge(TOP/BAR,1]state is TOP=[BAR=1]
* * *

Examples of failure scenarios

* *

Conflict of two transactions

* * This example illustrates two concurrent transactions, which derived from * same initial state of data tree and proposes conflicting modifications. * *
     * txA = broker.newWriteTransaction(); // allocates new transaction, data tree is empty
     * txB = broker.newWriteTransaction(); // allocates new transaction, data tree is empty
     *
     * txA.put(CONFIGURATION, PATH, A);    // writes to PATH value A
     * txB.put(CONFIGURATION, PATH, B)     // writes to PATH value B
     *
     * ListenableFuture futureA = txA.commit(); // transaction A is sealed and committed
     * ListenebleFuture futureB = txB.commit(); // transaction B is sealed and committed
     * 
* * Commit of transaction A will be processed asynchronously and data tree * will be updated to contain value A for PATH. * Returned {@link ListenableFuture} will successfully complete once * state is applied to data tree. * * Commit of Transaction B will fail, because previous transaction also * modified path in a concurrent way. The state introduced by transaction B * will not be applied. Returned {@link ListenableFuture} object will fail * with {@link OptimisticLockFailedException} exception, which indicates to * client that concurrent transaction prevented the submitted transaction from being * applied. * * @return Result of the Commit, containing success information or list of * encountered errors, if commit was not successful. The Future * blocks until {@link TransactionStatus#COMMITED} is reached. * Future will fail with {@link TransactionCommitFailedException} if * Commit of this transaction failed. TODO: Usability: Consider * change from ListenableFuture to * {@link com.google.common.util.concurrent.CheckedFuture} which * will throw {@link TransactionCommitFailedException}. * * @throws IllegalStateException * if the transaction is not {@link TransactionStatus#NEW} */ public ListenableFuture> commit(); }