About Concurrency Control in Greenplum Database
About Concurrency Control in Greenplum Database
Greenplum Database uses the PostgreSQL Multiversion Concurrency Control (MVCC) model to manage concurrent transactions for heap tables.
Concurrency control in a database management system allows concurrent queries to complete with correct results while ensuring the integrity of the database. Traditional databases use a two-phase locking protocol that prevents a transaction from modifying data that has been read by another concurrent transaction and prevents any concurrent transaction from reading or writing data that another transaction has updated. The locks required to coordinate transactions add contention to the database, reducing overall transaction throughput.
Greenplum Database uses the PostgreSQL Multiversion Concurrency Control (MVCC) model to manage concurrency for heap tables. With MVCC, each query operates on a snapshot of the database when the query starts. While it executes, a query cannot see changes made by other concurrent transactions. This ensures that a query sees a consistent view of the database. Queries that read rows can never block waiting for transactions that write rows. Conversely, queries that write rows cannot be blocked by transactions that read rows. This allows much greater concurrency than traditional database systems that employ locks to coordinate access between transactions that read and write data.
The MVCC model depends on the system's ability to manage multiple versions of data rows. A query operates on a snapshot of the database at the start of the query. A snapshot is the set of rows that are visible at the beginning of a statement or transaction. The snapshot ensures the query has a consistent and valid view of the database for the duration of its execution.
Each transaction is assigned a unique transaction ID (XID), an incrementing 32-bit value. When a new transaction starts, it is assigned the next XID. An SQL statement that is not enclosed in a transaction is treated as a single-statement transaction—the BEGIN and COMMIT are added implicitly. This is similar to autocommit in some database systems.
When a transaction inserts a row, the XID is saved with the row in the xmin system column. When a transaction deletes a row, the XID is saved in the xmax system column. Updating a row is treated as a delete and an insert, so the XID is saved to the xmax of the current row and the xmin of the newly inserted row. The xmin and xmax columns, together with the transaction completion status, specify a range of transactions for which the version of the row is visible. A transaction can see the effects of all transactions less than xmin, which are guaranteed to be committed, but it cannot see the effects of any transaction greater than or equal to xmax.
Multi-statement transactions must also record which command within a transaction inserted a row (cmin) or deleted a row (cmax) so that the transaction can see changes made by previous commands in the transaction. The command sequence is only relevant during the transaction, so the sequence is reset to 0 at the beginning of a transaction.
XID is a property of the database. Each segment database has its own XID sequence that cannot be compared to the XIDs of other segment databases. The master coordinates distributed transactions with the segments using a cluster-wide session ID number, called gp_session_id. The segments maintain a mapping of distributed transaction IDs with their local XIDs. The master coordinates distributed transactions across all of the segment with the two-phase commit protocol. If a transaction fails on any one segment, it is rolled back on all segments.
SELECT xmin, xmax, cmin, cmax, * FROM tablename;
Because you run the SELECT command on the master, the XIDs are the distributed transactions IDs. If you could execute the command in an individual segment database, the xmin and xmax values would be the segment's local XIDs.
Transaction ID Wraparound
The MVCC model uses transaction IDs (XIDs) to determine which rows are visible at the beginning of a query or transaction. The XID is a 32-bit value, so a database could theoretically execute over four billion transactions before the value overflows and wraps to zero. However, Greenplum Database uses modulo 232 arithmetic with XIDs, which allows the transaction IDs to wrap around, much as a clock wraps at twelve o'clock. For any given XID, there could be about two billion past XIDs and two billion future XIDs. This works until a version of a row persists through about two billion transactions, when it suddenly appears to be a new row. To prevent this, Greenplum has a special XID, called FrozenXID, which is always considered older than any regular XID it is compared with. The xmin of a row must be replaced with FrozenXID within two billion transactions, and this is one of the functions the VACUUM command performs.
Vacuuming the database at least every two billion transactions prevents XID wraparound. Greenplum Database monitors the transaction ID and warns if a VACUUM operation is required.
WARNING: database "database_name" must be vacuumed within number_of_transactions transactions
When the warning is issued, a VACUUM operation is required. If a VACUUM operation is not performed, Greenplum Database stops creating transactions to avoid possible data loss when it reaches a limit prior to when transaction ID wraparound occurs and issues this error:
FATAL: database is not accepting commands to avoid wraparound data loss in database "database_name"
See Recovering from a Transaction ID Limit Error for the procedure to recover from this error.
The server configuration parameters xid_warn_limit and xid_stop_limit control when the warning and error are displayed. The xid_warn_limit parameter specifies the number of transaction IDs before the xid_stop_limit when the warning is issued. The xid_stop_limit parameter specifies the number of transaction IDs before wraparound would occur when the error is issued and new transactions cannot be created.
Transaction Isolation Modes
- Dirty read – a transaction can read uncommitted data from another concurrent transaction.
- Non-repeatable read – a row read twice in a transaction can change because another concurrent transaction committed changes after the transaction began.
- Phantom read – a query executed twice in the same transaction can return two different sets of rows because another concurrent transaction added rows.
The SQL standard defines four transaction isolation modes that database systems must support:
|Level||Dirty Read||Non-Repeatable||Phantom Read|
The Greenplum Database SQL commands allow you to request READ UNCOMMITTED, READ COMMITTED, or SERIALIZABLE. Greenplum Database treats READ UNCOMMITTED the same as READ COMMITTED. Requesting REPEATABLE READ produces an error; use SERIALIZABLE instead. The default isolation mode is READ COMMITTED.
The difference between READ COMMITTED and SERIALIZABLE is that in READ COMMITTED mode, each statement in a transaction sees only rows committed before the statement started, while in SERIALIZABLE mode, all statements in a transaction see only rows committed before the transaction started.
The READ COMMITTED isolation mode permits greater concurrency and better performance than the SERIALIZABLE mode. It allows non-repeatable reads, where the values in a row retrieved twice in a transaction can differ because another concurrent transaction has committed changes since the transaction began. READ COMMITTED mode also permits phantom reads, where a query executed twice in the same transaction can return two different sets of rows.
The SERIALIZABLE isolation mode prevents both non-repeatable reads and phantom reads, but at the cost of concurrency and performance. Each concurrent transaction has a consistent view of the database taken at the beginning of execution. A concurrent transaction that attempts to modify data modified by another transaction is rolled back. Applications that execute transactions in SERIALIZABLE mode must be prepared to handle transactions that fail due to serialization errors. If SERIALIZABLE isolation mode is not required by the application, it is better to use READ COMMITTED mode.
The SQL standard specifies that concurrent serializable transactions produce the same database state they would produce if executed sequentially. The MVCC snapshot isolation model prevents dirty reads, non-repeatable reads, and phantom reads without expensive locking, but there are other interactions that can occur between some SERIALIZABLE transactions in Greenplum Database that prevent them from being truly serializable. These anomalies can often be attributed to the fact that Greenplum Database does not perform predicate locking, which means that a write in one transaction can affect the result of a previous read in another concurrent transaction.
Transactions that run concurrently should be examined to identify interactions that are not prevented by disallowing concurrent updates of the same data. Problems identified can be prevented by using explicit table locks or by requiring the conflicting transactions to update a dummy row introduced to represent the conflict.
BEGIN; SET TRANSACTION ISOLATION LEVEL SERIALIZABLE; ... COMMIT;
BEGIN TRANSACTION ISOLATION LEVEL SERIALIZABLE;
The default transaction isolation mode can be changed for a session by setting the default_transaction_isolation configuration property.
Removing Dead Rows from Tables
Updating or deleting a row leaves an expired version of the row in the table. When an expired row is no longer referenced by any active transactions, it can be removed and the space it occupied can be reused. The VACUUM command marks the space used by expired rows for reuse.
When expired rows accumulate in a table, the disk files must be extended to accommodate new rows. Performance suffers due to the increased disk I/O required to execute queries. This condition is called bloat and it should be managed by regularly vacuuming tables.
The VACUUM command (without FULL) can run concurrently with other queries. It marks the space previously used by the expired rows as free. If the amount of remaining free space is significant, it adds the page to the table's free space map. When Greenplum Database later needs space for new rows, it first consults the table's free space map to find pages with available space. If none are found, new pages will be appended to the file.
VACUUM (without FULL) does not consolidate pages or reduce the size of the table on disk. The space it recovers is only available through the free space map. To prevent disk files from growing, it is important to run VACUUM often enough. The frequency of required VACUUM runs depends on the frequency of updates and deletes in the table (inserts only ever add new rows). Heavily updated tables might require several VACUUM runs per day, to ensure that the available free space can be found through the free space map. It is also important to run VACUUM after running a transaction that updates or deletes a large number of rows.
The VACUUM FULL command rewrites the table without expired rows, reducing the table to its minimum size. Every page in the table is checked, and visible rows are moved up into pages which are not yet fully packed. Empty pages are discarded. The table is locked until VACUUM FULL completes. This is very expensive compared to the regular VACUUM command, and can be avoided or postponed by vacuuming regularly. It is best to run VACUUM FULL during a maintenance period. An alternative to VACUUM FULL is to recreate the table with a CREATE TABLE AS statement and then drop the old table.
The free space map resides in shared memory and keeps track of free space for all tables and indexes. Each table or index uses about 60 bytes of memory and each page with free space consumes six bytes. Two system configuration parameters configure the size of the free space map:
- Sets the maximum number of disk pages that can be added to the shared free space map. Six bytes of shared memory are consumed for each page slot. The default is 200000. This parameter must be set to at least 16 times the value of max_fsm_relations.
Sets the maximum number of relations that will be tracked in the shared memory free space map. This parameter should be set to a value larger than the total number of tables + indexes + system tables. The default is 1000. About 60 bytes of memory are consumed for each relation per segment instance. It is better to set the parameter too high than too low.
If the free space map is undersized, some disk pages with available space will not be added to the map, and that space cannot be reused until at least the next VACUUM command runs. This causes files to grow.
You can run VACUUM VERBOSE tablename to get a report, by segment, of the number of dead rows removed, the number of pages affected, and the number of pages with usable free space.
SELECT relname, relpages, reltuples FROM pg_class WHERE relname='tablename';
Another useful tool is the gp_bloat_diag view in the gp_toolkit schema, which identifies bloat in tables by comparing the actual number of pages used by a table to the expected number. See "The gp_toolkit Administrative Schema" in the Greenplum Database Reference Guide for more about gp_bloat_diag.