21.6.11 NDB Cluster Replication Conflict Resolution
When using a replication setup involving multiple sources (including circular replication), it is possible that different sources may try to update the same row on the replica with different data. Conflict resolution in NDB Cluster Replication provides a means of resolving such conflicts by permitting a user-defined resolution column to be used to determine whether or not an update on a given source should be applied on the replica.
Some types of conflict resolution supported by NDB Cluster (
NDB$MAX_DELETE_WIN()) implement this user-defined column as a “timestamp” column (although its type cannot be
TIMESTAMP, as explained later in this section). These types of conflict resolution are always applied a row-by-row basis rather than a transactional basis. The epoch-based conflict resolution functions
NDB$EPOCH_TRANS() compare the order in which epochs are replicated (and thus these functions are transactional). Different methods can be used to compare resolution column values on the replica when conflicts occur, as explained later in this section; the method used can be set on a per-table basis.
You should also keep in mind that it is the application's responsibility to ensure that the resolution column is correctly populated with relevant values, so that the resolution function can make the appropriate choice when determining whether to apply an update.
Requirements. Preparations for conflict resolution must be made on both the source and the replica. These tasks are described in the following list:
On the source writing the binary logs, you must determine which columns are sent (all columns or only those that have been updated). This is done for the MySQL Server as a whole by applying the mysqld startup option
--ndb-log-updated-only(described later in this section) or on a per-table basis by entries in the
mysql.ndb_replicationtable (see The ndb_replication system table).Note
If you are replicating tables with very large columns (such as
--ndb-log-updated-onlycan also be useful for reducing the size of the binary logs and avoiding possible replication failures due to exceeding
See Section 18.104.22.168, “Replication and max_allowed_packet”, for more information about this issue.
On the replica, you must determine which type of conflict resolution to apply (“latest timestamp wins”, “same timestamp wins”, “primary wins”, “primary wins, complete transaction”, or none). This is done using the
mysql.ndb_replicationsystem table, on a per-table basis (see The ndb_replication system table).
NDB Cluster also supports read conflict detection, that is, detecting conflicts between reads of a given row in one cluster and updates or deletes of the same row in another cluster. This requires exclusive read locks obtained by setting
ndb_log_exclusive_readsequal to 1 on the replica. All rows read by a conflicting read are logged in the exceptions table. For more information, see Read conflict detection and resolution.
When using the functions
NDB$MAX_DELETE_WIN() for timestamp-based conflict resolution, we often refer to the column used for determining updates as a “timestamp” column. However, the data type of this column is never
TIMESTAMP; instead, its data type should be
BIGINT. The “timestamp” column should also be
NDB$EPOCH_TRANS() functions discussed later in this section work by comparing the relative order of replication epochs applied on a primary and secondary NDB Cluster, and do not make use of timestamps.
Source column control. We can see update operations in terms of “before” and “after” images—that is, the states of the table before and after the update is applied. Normally, when updating a table with a primary key, the “before” image is not of great interest; however, when we need to determine on a per-update basis whether or not to use the updated values on a replica, we need to make sure that both images are written to the source's binary log. This is done with the
--ndb-log-update-as-write option for mysqld, as described later in this section.
Whether logging of complete rows or of updated columns only is done is decided when the MySQL server is started, and cannot be changed online; you must either restart mysqld, or start a new mysqld instance with different logging options.
Log complete rows
Log only column data that has been updated—that is, column data whose value has been set, regardless of whether or not this value was actually changed. This is the default behavior.
It is usually sufficient—and more efficient—to log updated columns only; however, if you need to log full rows, you can do so by setting
The setting of the MySQL Server's
--ndb-log-update-as-write option determines whether logging is performed with or without the “before” image. Because conflict resolution is done in the MySQL Server's update handler, it is necessary to control logging performed by the replication source such that updates are updates and not writes; that is, such that updates are treated as changes in existing rows rather than the writing of new rows, even though these replace existing rows. This option is turned on by default; in other words, updates are treated as writes. That is, updates are by default written as
write_row events in the binary log, rather than as
To disable the option, start the source mysqld with
--ndb-log-update-as-write=OFF. You must do this when replicating from NDB tables to tables using a different storage engine; see Replication from NDB to other storage engines, and Replication from NDB to a nontransactional storage engine, for more information.
Conflict resolution control. Conflict resolution is usually enabled on the server where conflicts can occur. Like logging method selection, it is enabled by entries in the
The ndb_replication system table. To enable conflict resolution, it is necessary to create an
ndb_replication table in the
mysql system database on the source, the replica, or both, depending on the conflict resolution type and method to be employed. This table is used to control logging and conflict resolution functions on a per-table basis, and has one row per table involved in replication.
ndb_replication is created and filled with control information on the server where the conflict is to be resolved. In a simple source-replica setup where data can also be changed locally on the replica this is typically the replica. In a more complex replication scheme, such as bidirectional replication, this is usually all of the sources involved. Each row in
mysql.ndb_replication corresponds to a table being replicated, and specifies how to log and resolve conflicts (that is, which conflict resolution function, if any, to use) for that table. The definition of the
mysql.ndb_replication table is shown here:
CREATE TABLE mysql.ndb_replication ( db VARBINARY(63), table_name VARBINARY(63), server_id INT UNSIGNED, binlog_type INT UNSIGNED, conflict_fn VARBINARY(128), PRIMARY KEY USING HASH (db, table_name, server_id) ) ENGINE=NDB PARTITION BY KEY(db,table_name);
The columns in this table are described in the next few paragraphs.
db. The name of the database containing the table to be replicated. You may employ either or both of the wildcards
% as part of the database name. Matching is similar to what is implemented for the
table_name. The name of the table to be replicated. The table name may include either or both of the wildcards
%. Matching is similar to what is implemented for the
server_id. The unique server ID of the MySQL instance (SQL node) where the table resides.
binlog_type. The type of binary logging to be employed. This is determined as shown in the following table:
Table 21.284 binlog_type values, with internal values and descriptions
||Use server default|
||Do not log this table in the binary log|
||Only updated attributes are logged|
||Log full row, even if not updated (MySQL server default behavior)|
||Use updated attributes, even if values are unchanged|
||Use full row, even if values are unchanged|
conflict_fn. The conflict resolution function to be applied. This function must be specified as one of those shown in the following list:
These functions are described in the next few paragraphs.
NDB$OLD(column_name). If the value of
column_name is the same on both the source and the replica, then the update is applied; otherwise, the update is not applied on the replica and an exception is written to the log. This is illustrated by the following pseudocode:
if (source_old_column_value == replica_current_column_value) apply_update(); else log_exception();
The column value from the source's “before” image is used by this function.
NDB$MAX(column_name). If the “timestamp” column value for a given row coming from the source is higher than that on the replica, it is applied; otherwise it is not applied on the replica. This is illustrated by the following pseudocode:
if (source_new_column_value > replica_current_column_value) apply_update();
This function can be used for “greatest timestamp wins” conflict resolution. This type of conflict resolution ensures that, in the event of a conflict, the version of the row that was most recently updated is the version that persists.
The column value from the sources's “after” image is used by this function.
NDB$MAX_DELETE_WIN(). This is a variation on
NDB$MAX(). Due to the fact that no timestamp is available for a delete operation, a delete using
NDB$MAX() is in fact processed as
NDB$OLD, but for some use cases, this is not optimal. For
NDB$MAX_DELETE_WIN(), if the “timestamp” column value for a given row adding or updating an existing row coming from the source is higher than that on the replica, it is applied. However, delete operations are treated as always having the higher value. This is illustrated by the following pseudocode:
if ( (source_new_column_value > replica_current_column_value) || operation.type == "delete") apply_update();
This function can be used for “greatest timestamp, delete wins” conflict resolution. This type of conflict resolution ensures that, in the event of a conflict, the version of the row that was deleted or (otherwise) most recently updated is the version that persists.
NDB$MAX(), the column value from the source's “after” image is the value used by this function.
NDB$EPOCH() and NDB$EPOCH_TRANS(). The
NDB$EPOCH() function tracks the order in which replicated epochs are applied on a replica cluster relative to changes originating on the replica. This relative ordering is used to determine whether changes originating on the replica are concurrent with any changes that originate locally, and are therefore potentially in conflict.
Most of what follows in the description of
NDB$EPOCH() also applies to
NDB$EPOCH_TRANS(). Any exceptions are noted in the text.
NDB$EPOCH() is asymmetric, operating on one NDB Cluster in a bidirectional replication configuration (sometimes referred to as “active-active” replication). We refer here to cluster on which it operates as the primary, and the other as the secondary. The replica on the primary is responsible for detecting and handling conflicts, while the replica on the secondary is not involved in any conflict detection or handling.
When the replica on the primary detects conflicts, it injects events into its own binary log to compensate for these; this ensures that the secondary NDB Cluster eventually realigns itself with the primary and so keeps the primary and secondary from diverging. This compensation and realignment mechanism requires that the primary NDB Cluster always wins any conflicts with the secondary—that is, that the primary's changes are always used rather than those from the secondary in event of a conflict. This “primary always wins” rule has the following implications:
Operations that change data, once committed on the primary, are fully persistent and are not undone or rolled back by conflict detection and resolution.
Data read from the primary is fully consistent. Any changes committed on the Primary (locally or from the replica) are not reverted later.
Operations that change data on the secondary may later be reverted if the primary determines that they are in conflict.
Individual rows read on the secondary are self-consistent at all times, each row always reflecting either a state committed by the secondary, or one committed by the primary.
Sets of rows read on the secondary may not necessarily be consistent at a given single point in time. For
NDB$EPOCH_TRANS(), this is a transient state; for
NDB$EPOCH(), it can be a persistent state.
Assuming a period of sufficient length without any conflicts, all data on the secondary NDB Cluster (eventually) becomes consistent with the primary's data.
NDB$EPOCH_TRANS() do not require any user schema modifications, or application changes to provide conflict detection. However, careful thought must be given to the schema used, and the access patterns used, to verify that the complete system behaves within specified limits.
Each of the
NDB$EPOCH_TRANS() functions can take an optional parameter; this is the number of bits to use to represent the lower 32 bits of the epoch, and should be set to no less than the value calculated as shown here:
CEIL( LOG2( TimeBetweenGlobalCheckpoints / TimeBetweenEpochs ), 1)
For the default values of these configuration parameters (2000 and 100 milliseconds, respectively), this gives a value of 5 bits, so the default value (6) should be sufficient, unless other values are used for
TimeBetweenEpochs, or both. A value that is too small can result in false positives, while one that is too large could lead to excessive wasted space in the database.
NDB$EPOCH_TRANS() insert entries for conflicting rows into the relevant exceptions tables, provided that these tables have been defined according to the same exceptions table schema rules as described elsewhere in this section (see NDB$OLD(column_name)). You must create any exceptions table before creating the data table with which it is to be used.
As with the other conflict detection functions discussed in this section,
NDB$EPOCH_TRANS() are activated by including relevant entries in the
mysql.ndb_replication table (see The ndb_replication system table). The roles of the primary and secondary NDB Clusters in this scenario are fully determined by
mysql.ndb_replication table entries.
Because the conflict detection algorithms employed by
NDB$EPOCH_TRANS() are asymmetric, you must use different values for the
server_id entries of the primary and secondary replicas.
A conflict between
DELETE operations alone is not sufficient to trigger a conflict using
NDB$EPOCH_TRANS(), and the relative placement within epochs does not matter. (Bug #18459944)
Conflict detection status variables. Several status variables can be used to monitor conflict detection. You can see how many rows have been found in conflict by
NDB$EPOCH() since this replica was last restarted from the current value of the
Ndb_conflict_fn_epoch system status variable.
Ndb_conflict_fn_epoch_trans provides the number of rows that have been found directly in conflict by
Ndb_conflict_fn_epoch2_trans show the number of rows found in conflict by
NDB$EPOCH2_TRANS(), respectively. The number of rows actually realigned, including those affected due to their membership in or dependency on the same transactions as other conflicting rows, is given by
For more information, see Section 22.214.171.124.3, “NDB Cluster Status Variables”.
Conflicts are detected using NDB Cluster epoch boundaries, with granularity proportional to
TimeBetweenEpochs(default: 100 milliseconds). The minimum conflict window is the minimum time during which concurrent updates to the same data on both clusters always report a conflict. This is always a nonzero length of time, and is roughly proportional to
2 * (latency + queueing + TimeBetweenEpochs). This implies that—assuming the default for
TimeBetweenEpochsand ignoring any latency between clusters (as well as any queuing delays)—the minimum conflict window size is approximately 200 milliseconds. This minimum window should be considered when looking at expected application “race” patterns.
Additional storage is required for tables using the
NDB$EPOCH_TRANS()functions; from 1 to 32 bits extra space per row is required, depending on the value passed to the function.
Conflicts between delete operations may result in divergence between the primary and secondary. When a row is deleted on both clusters concurrently, the conflict can be detected, but is not recorded, since the row is deleted. This means that further conflicts during the propagation of any subsequent realignment operations are not detected, which can lead to divergence.
Deletes should be externally serialized, or routed to one cluster only. Alternatively, a separate row should be updated transactionally with such deletes and any inserts that follow them, so that conflicts can be tracked across row deletes. This may require changes in applications.
Only two NDB Clusters in a birectional “active-active” configuration are currently supported when using
NDB$EPOCH_TRANS()for conflict detection.
NDB$EPOCH_TRANS() extends the
NDB$EPOCH() function. Conflicts are detected and handled in the same way using the “primary wins all” rule (see NDB$EPOCH() and NDB$EPOCH_TRANS()) but with the extra condition that any other rows updated in the same transaction in which the conflict occurred are also regarded as being in conflict. In other words, where
NDB$EPOCH() realigns individual conflicting rows on the secondary,
NDB$EPOCH_TRANS() realigns conflicting transactions.
In addition, any transactions which are detectably dependent on a conflicting transaction are also regarded as being in conflict, these dependencies being determined by the contents of the secondary cluster's binary log. Since the binary log contains only data modification operations (inserts, updates, and deletes), only overlapping data modifications are used to determine dependencies between transactions.
NDB$EPOCH_TRANS() is subject to the same conditions and limitations as
NDB$EPOCH(), and in addition requires that Version 2 binary log row events are used (
log_bin_use_v1_row_events equal to 0), which adds a storage overhead of 2 bytes per event in the binary log. In addition, all transaction IDs must be recorded in the secondary's binary log (
--ndb-log-transaction-id option), which adds a further variable overhead (up to 13 bytes per row).
Status information. A server status variable
Ndb_conflict_fn_max provides a count of the number of times that a row was not applied on the current SQL node due to “greatest timestamp wins” conflict resolution since the last time that mysqld was started.
The number of times that a row was not applied as the result of “same timestamp wins” conflict resolution on a given mysqld since the last time it was restarted is given by the global status variable
Ndb_conflict_fn_old. In addition to incrementing
Ndb_conflict_fn_old, the primary key of the row that was not used is inserted into an exceptions table, as explained later in this section.
NDB$EPOCH2() function is similar to
NDB$EPOCH(), except that
NDB$EPOCH2() provides for delete-delete handling with a bidirectional replication topology. In this scenario, primary and secondary roles are assigned to the two sources by setting the
ndb_slave_conflict_role system variable to the appropriate value on each source (usually one each of
SECONDARY). When this is done, modifications made by the secondary are reflected by the primary back to the secondary which then conditionally applies them.
NDB$EPOCH2_TRANS() extends the
NDB$EPOCH2() function. Conflicts are detected and handled in the same way, and assigning primary and secondary roles to the replicating clusters, but with the extra condition that any other rows updated in the same transaction in which the conflict occurred are also regarded as being in conflict. That is,
NDB$EPOCH2() realigns individual conflicting rows on the secondary, while
NDB$EPOCH_TRANS() realigns conflicting transactions.
NDB$EPOCH_TRANS() use metadata that is specified per row, per last modified epoch, to determine on the primary whether an incoming replicated row change from the secondary is concurrent with a locally committed change; concurrent changes are regarded as conflicting, with subesequent exceptions table updates and realignment of the secondary. A problem arises when a row is deleted on the primary so there is no longer any last-modified epoch available to determine whether any replicated operations conflict, which means that conflicting delete operationss are not detected. This can result in divergence, an example being a delete on one cluster which is concurrent with a delete and insert on the other; this why delete operations can be routed to only one cluster when using
NDB$EPOCH2() bypasses the issue just described—storing information about deleted rows on the PRIMARY—by ignoring any delete-delete conflict, and by avoiding any potential resultant divergence as well. This is accomplished by reflecting any operation successfully applied on and replicated from the secondary back to the secondary. On its return to the secondary, it can be used to reapply an operation on the secondary which was deleted by an operation originating from the primary.
NDB$EPOCH2(), you should keep in mind that the secondary applies the delete from the primary, removing the new row until it is restored by a reflected operation. In theory, the subsequent insert or update on the secondary conflicts with the delete from the primary, but in this case, we choose to ignore this and allow the secondary to “win”, in the interest of preventing divergence between the clusters. In other words, after a delete, the primary does not detect conflicts, and instead adopts the secondary's following changes immediately. Because of this, the secondary's state can revisit multiple previous committed states as it progresses to a final (stable) state, and some of these may be visible.
You should also be aware that reflecting all operations from the secondary back to the primary increases the size of the primary's logbinary log, as well as demands on bandwidth, CPU usage, and disk I/O.
Application of reflected operations on the secondary depends on the state of the target row on the secondary. Whether or not reflected changes are applied on the secondary can be tracked by checking the
Ndb_conflict_reflected_op_discard_count status variables. The number of changes applied is simply the difference between these two values (note that
Ndb_conflict_reflected_op_prepare_count is always greater than or equal to
Events are applied if and only if both of the following conditions are true:
The existence of the row—that is, whether or not it exists—is in accordance with the type of event. For delete and update operations, the row must already exist. For insert operations, the row must not exist.
The row was last modified by the primary. It is possible that the modification was accomplished through the execution of a reflected operation.
If both of the conditions are not met, the reflected operation is discarded by the secondary.
Conflict resolution exceptions table. To use the
NDB$OLD() conflict resolution function, it is also necessary to create an exceptions table corresponding to each
NDB table for which this type of conflict resolution is to be employed. This is also true when using
NDB$EPOCH_TRANS(). The name of this table is that of the table for which conflict resolution is to be applied, with the string
$EX appended. (For example, if the name of the original table is
mytable, the name of the corresponding exceptions table name should be
mytable$EX.) The syntax for creating the exceptions table is as shown here:
CREATE TABLE original_table$EX ( [NDB$]server_id INT UNSIGNED, [NDB$]source_server_id INT UNSIGNED, [NDB$]source_epoch BIGINT UNSIGNED, [NDB$]count INT UNSIGNED, [NDB$OP_TYPE ENUM('WRITE_ROW','UPDATE_ROW', 'DELETE_ROW', 'REFRESH_ROW', 'READ_ROW') NOT NULL,] [NDB$CFT_CAUSE ENUM('ROW_DOES_NOT_EXIST', 'ROW_ALREADY_EXISTS', 'DATA_IN_CONFLICT', 'TRANS_IN_CONFLICT') NOT NULL,] [NDB$ORIG_TRANSID BIGINT UNSIGNED NOT NULL,] original_table_pk_columns, [orig_table_column|orig_table_column$OLD|orig_table_column$NEW,] [additional_columns,] PRIMARY KEY([NDB$]server_id, [NDB$]source_server_id, [NDB$]source_epoch, [NDB$]count) ) ENGINE=NDB;
The first four columns are required. The names of the first four columns and the columns matching the original table's primary key columns are not critical; however, we suggest for reasons of clarity and consistency, that you use the names shown here for the
count columns, and that you use the same names as in the original table for the columns matching those in the original table's primary key.
If the exceptions table uses one or more of the optional columns
NDB$ORIG_TRANSID discussed later in this section, then each of the required columns must also be named using the prefix
NDB$. If desired, you can use the
NDB$ prefix to name the required columns even if you do not define any optional columns, but in this case, all four of the required columns must be named using the prefix.
Following these columns, the columns making up the original table's primary key should be copied in the order in which they are used to define the primary key of the original table. The data types for the columns duplicating the primary key columns of the original table should be the same as (or larger than) those of the original columns. A subset of the primary key columns may be used.
The exceptions table must use the
NDB storage engine. (An example that uses
NDB$OLD() with an exceptions table is shown later in this section.)
Additional columns may optionally be defined following the copied primary key columns, but not before any of them; any such extra columns cannot be
NOT NULL. NDB Cluster supports three additional, predefined optional columns
NDB$ORIG_TRANSID, which are described in the next few paragraphs.
NDB$OP_TYPE ENUM('WRITE_ROW', 'UPDATE_ROW', 'DELETE_ROW', 'REFRESH_ROW', 'READ_ROW') NOT NULL
DELETE_ROW operation types represent user-initiated operations.
REFRESH_ROW operations are operations generated by conflict resolution in compensating transactions sent back to the originating cluster from the cluster that detected the conflict.
READ_ROW operations are user-initiated read tracking operations defined with exclusive row locks.
NDB$CFT_CAUSE ENUM('ROW_DOES_NOT_EXIST', 'ROW_ALREADY_EXISTS', 'DATA_IN_CONFLICT', 'TRANS_IN_CONFLICT') NOT NULL
ROW_DOES_NOT_EXIST can be reported as the cause for
ROW_ALREADY_EXISTS can be reported for
DATA_IN_CONFLICT is reported when a row-based conflict function detects a conflict;
TRANS_IN_CONFLICT is reported when a transactional conflict function rejects all of the operations belonging to a complete transaction.
NDB$ORIG_TRANSID BIGINT UNSIGNED NOT NULL
NDB$ORIG_TRANSID is a 64-bit value generated by
NDB. This value can be used to correlate multiple exceptions table entries belonging to the same conflicting transaction from the same or different exceptions tables.
Additional reference columns which are not part of the original table's primary key can be named
references old values in update and delete operations—that is, operations containing
can be used to reference new values in insert and update operations—in other words, operations using
UPDATE_ROW events, or both types of events. Where a conflicting operation does not supply a value for a given reference column that is not a primary key, the exceptions table row contains either
NULL, or a defined default value for that column.
mysql.ndb_replication table is read when a data table is set up for replication, so the row corresponding to a table to be replicated must be inserted into
mysql.ndb_replication before the table to be replicated is created.
The following examples assume that you have already a working NDB Cluster replication setup, as described in Section 21.6.5, “Preparing the NDB Cluster for Replication”, and Section 21.6.6, “Starting NDB Cluster Replication (Single Replication Channel)”.
On the source, perform this
INSERT INTO mysql.ndb_replication VALUES ('test', 't1', 0, NULL, 'NDB$MAX(mycol)');
Inserting a 0 into the
server_idindicates that all SQL nodes accessing this table should use conflict resolution. If you want to use conflict resolution on a specific mysqld only, use the actual server ID.
binlog_typecolumn has the same effect as inserting 0 (
NBT_DEFAULT); the server default is used.
CREATE TABLE test.t1 ( columns mycol INT UNSIGNED, columns ) ENGINE=NDB;
Now, when updates are performed on this table, conflict resolution is applied, and the version of the row having the greatest value for
mycolis written to the replica.
binlog_type options—such as
NBT_UPDATED_ONLY_USE_UPDATE should be used to control logging on the source using the
ndb_replication table rather than by using command-line options.
NDB$OLD() example. Suppose an
NDB table such as the one defined here is being replicated, and you wish to enable “same timestamp wins” conflict resolution for updates to this table:
CREATE TABLE test.t2 ( a INT UNSIGNED NOT NULL, b CHAR(25) NOT NULL, columns, mycol INT UNSIGNED NOT NULL, columns, PRIMARY KEY pk (a, b) ) ENGINE=NDB;
The following steps are required, in the order shown:
First—and prior to creating
test.t2—you must insert a row into the
mysql.ndb_replicationtable, as shown here:
INSERT INTO mysql.ndb_replication VALUES ('test', 't2', 0, NULL, 'NDB$OLD(mycol)');
Possible values for the
binlog_typecolumn are shown earlier in this section. The value
'NDB$OLD(mycol)'should be inserted into the
Create an appropriate exceptions table for
test.t2. The table creation statement shown here includes all required columns; any additional columns must be declared following these columns, and before the definition of the table's primary key.
CREATE TABLE test.t2$EX ( server_id INT UNSIGNED, source_server_id INT UNSIGNED, source_epoch BIGINT UNSIGNED, count INT UNSIGNED, a INT UNSIGNED NOT NULL, b CHAR(25) NOT NULL, [additional_columns,] PRIMARY KEY(server_id, source_server_id, source_epoch, count) ) ENGINE=NDB;
We can include additional columns for information about the type, cause, and originating transaction ID for a given conflict. We are also not required to supply matching columns for all primary key columns in the original table. This means you can create the exceptions table like this:
CREATE TABLE test.t2$EX ( NDB$server_id INT UNSIGNED, NDB$source_server_id INT UNSIGNED, NDB$source_epoch BIGINT UNSIGNED, NDB$count INT UNSIGNED, a INT UNSIGNED NOT NULL, NDB$OP_TYPE ENUM('WRITE_ROW','UPDATE_ROW', 'DELETE_ROW', 'REFRESH_ROW', 'READ_ROW') NOT NULL, NDB$CFT_CAUSE ENUM('ROW_DOES_NOT_EXIST', 'ROW_ALREADY_EXISTS', 'DATA_IN_CONFLICT', 'TRANS_IN_CONFLICT') NOT NULL, NDB$ORIG_TRANSID BIGINT UNSIGNED NOT NULL, [additional_columns,] PRIMARY KEY(NDB$server_id, NDB$source_server_id, NDB$source_epoch, NDB$count) ) ENGINE=NDB;Note
NDB$prefix is required for the four required columns since we included at least one of the columns
NDB$ORIG_TRANSIDin the table definition.
Create the table
test.t2as shown previously.
These steps must be followed for every table for which you wish to perform conflict resolution using
NDB$OLD(). For each such table, there must be a corresponding row in
mysql.ndb_replication, and there must be an exceptions table in the same database as the table being replicated.
Read conflict detection and resolution. NDB Cluster also supports tracking of read operations, which makes it possible in circular replication setups to manage conflicts between reads of a given row in one cluster and updates or deletes of the same row in another. This example uses
department tables to model a scenario in which an employee is moved from one department to another on the source cluster (which we refer to hereafter as cluster A) while the replica cluster (hereafter B) updates the employee count of the employee's former department in an interleaved transaction.
The data tables have been created using the following SQL statements:
# Employee table CREATE TABLE employee ( id INT PRIMARY KEY, name VARCHAR(2000), dept INT NOT NULL ) ENGINE=NDB; # Department table CREATE TABLE department ( id INT PRIMARY KEY, name VARCHAR(2000), members INT ) ENGINE=NDB;
The contents of the two tables include the rows shown in the (partial) output of the following
mysql> SELECT id, name, dept FROM employee; +---------------+------+ | id | name | dept | +------+--------+------+ ... | 998 | Mike | 3 | | 999 | Joe | 3 | | 1000 | Mary | 3 | ... +------+--------+------+ mysql> SELECT id, name, members FROM department; +-----+-------------+---------+ | id | name | members | +-----+-------------+---------+ ... | 3 | Old project | 24 | ... +-----+-------------+---------+
We assume that we are already using an exceptions table that includes the four required columns (and these are used for this table's primary key), the optional columns for operation type and cause, and the original table's primary key column, created using the SQL statement shown here:
CREATE TABLE employee$EX ( NDB$server_id INT UNSIGNED, NDB$source_server_id INT UNSIGNED, NDB$source_epoch BIGINT UNSIGNED, NDB$count INT UNSIGNED, NDB$OP_TYPE ENUM( 'WRITE_ROW','UPDATE_ROW', 'DELETE_ROW', 'REFRESH_ROW','READ_ROW') NOT NULL, NDB$CFT_CAUSE ENUM( 'ROW_DOES_NOT_EXIST', 'ROW_ALREADY_EXISTS', 'DATA_IN_CONFLICT', 'TRANS_IN_CONFLICT') NOT NULL, id INT NOT NULL, PRIMARY KEY(NDB$server_id, NDB$source_server_id, NDB$source_epoch, NDB$count) ) ENGINE=NDB;
Suppose there occur the two simultaneous transactions on the two clusters. On cluster A, we create a new department, then move employee number 999 into that department, using the following SQL statements:
BEGIN; INSERT INTO department VALUES (4, "New project", 1); UPDATE employee SET dept = 4 WHERE id = 999; COMMIT;
At the same time, on cluster B, another transaction reads from
employee, as shown here:
BEGIN; SELECT name FROM employee WHERE id = 999; UPDATE department SET members = members - 1 WHERE id = 3; commit;
The conflicting transactions are not normally detected by the conflict resolution mechanism, since the conflict is between a read (
SELECT) and an update operation. You can circumvent this issue by executing
= 1 on the replica cluster. Acquiring exclusive read locks in this way causes any rows read on the source to be flagged as needing conflict resolution on the replica cluster. If we enable exclusive reads in this way prior to the logging of these transactions, the read on cluster B is tracked and sent to cluster A for resolution; the conflict on the employee row is subsequently detected and the transaction on cluster B is aborted.
The conflict is registered in the exceptions table (on cluster A) as a
READ_ROW operation (see Conflict resolution exceptions table, for a description of operation types), as shown here:
mysql> SELECT id, NDB$OP_TYPE, NDB$CFT_CAUSE FROM employee$EX; +-------+-------------+-------------------+ | id | NDB$OP_TYPE | NDB$CFT_CAUSE | +-------+-------------+-------------------+ ... | 999 | READ_ROW | TRANS_IN_CONFLICT | +-------+-------------+-------------------+
Any existing rows found in the read operation are flagged. This means that multiple rows resulting from the same conflict may be logged in the exception table, as shown by examining the effects a conflict between an update on cluster A and a read of multiple rows on cluster B from the same table in simultaneous transactions. The transaction executed on cluster A is shown here:
BEGIN; INSERT INTO department VALUES (4, "New project", 0); UPDATE employee SET dept = 4 WHERE dept = 3; SELECT COUNT(*) INTO @count FROM employee WHERE dept = 4; UPDATE department SET members = @count WHERE id = 4; COMMIT;
Concurrently a transaction containing the statements shown here runs on cluster B:
SET ndb_log_exclusive_reads = 1; # Must be set if not already enabled ... BEGIN; SELECT COUNT(*) INTO @count FROM employee WHERE dept = 3 FOR UPDATE; UPDATE department SET members = @count WHERE id = 3; COMMIT;
In this case, all three rows matching the
WHERE condition in the second transaction's
SELECT are read, and are thus flagged in the exceptions table, as shown here:
mysql> SELECT id, NDB$OP_TYPE, NDB$CFT_CAUSE FROM employee$EX; +-------+-------------+-------------------+ | id | NDB$OP_TYPE | NDB$CFT_CAUSE | +-------+-------------+-------------------+ ... | 998 | READ_ROW | TRANS_IN_CONFLICT | | 999 | READ_ROW | TRANS_IN_CONFLICT | | 1000 | READ_ROW | TRANS_IN_CONFLICT | ... +-------+-------------+-------------------+
Read tracking is performed on the basis of existing rows only. A read based on a given condition track conflicts only of any rows that are found and not of any rows that are inserted in an interleaved transaction. This is similar to how exclusive row locking is performed in a single instance of NDB Cluster.