mirror of
https://github.com/timescale/timescaledb.git
synced 2025-05-22 05:31:41 +08:00
ffdc095d6e
Currently CREATE INDEX creates the indices for all chunks in a single
transaction, which holds a lock on the root hypertable and all chunks. This
means that during CREATE INDEX no new inserts can occur, even if we're not
currently building an index on the table being inserted to.
This commit adds the option to create indices using a separate
transaction for each chunk. This option, used like
CREATE INDEX ON <table> WITH (timescaledb.transaction_per_chunk);
should cause less contention than a regular CREATE INDEX, in exchange
for the possibility that the index will be created on only some, or none,
of the chunks, if the command fails partway through. The command holds a lock on
the root index used as a template throughout the command, and each of the only
additionally locks the chunk being indexed. This means that that chunks which
are not currently being indexed can be inserted to, and new chunks can be
created while the CREATE INDEX command is in progress.
To enable detection of failed transaction_per_chunk CREATE INDEXs, the
hypertable's index is marked as invalid while the CREATE INDEX is in progress,
if the command fails partway through, the index will remain invalid. If such an
invalid index is discovered, it can be dropped an recreated to ensure that all
chunks have a copy of the index, in the future, we may add a command to create
indexes on only those chunks which are missing them. Note that even though the
hypertable's index is marked as invalid, new chunks will have a copy of the
index build as normal.
As part of the refactoring to make this command work, normal index creation was
slightly modified. Instead of getting the column names an index uses
one-at-a-time we get them all at once at the beginning of index creation, this
allows to close the hypertable's root table once we've determined all of while
we create the index info for each chunk. Secondly, it changes our function to
lookup a tablespace, ts_hypertable_get_tablespace_at_offset_from, to only take a
hypertable id, instead of the hypertable's entire cache entry; this function
only ever used the id, so this allows us to release the hypertable cache earlier
145 lines
5.7 KiB
PL/PgSQL
145 lines
5.7 KiB
PL/PgSQL
-- This file and its contents are licensed under the Apache License 2.0.
|
|
-- Please see the included NOTICE for copyright information and
|
|
-- LICENSE-APACHE for a copy of the license.
|
|
|
|
CREATE TABLE index_test(id serial, time timestamptz, device integer, temp float);
|
|
SELECT * FROM test.show_columns('index_test');
|
|
|
|
-- Test that we can handle difference in attnos across hypertable and
|
|
-- chunks by dropping the ID column
|
|
ALTER TABLE index_test DROP COLUMN id;
|
|
SELECT * FROM test.show_columns('index_test');
|
|
|
|
-- No pre-existing UNIQUE index, so partitioning on two columns should work
|
|
SELECT create_hypertable('index_test', 'time', 'device', 2);
|
|
|
|
INSERT INTO index_test VALUES ('2017-01-20T09:00:01', 1, 17.5);
|
|
|
|
\set ON_ERROR_STOP 0
|
|
-- cannot create a UNIQUE index with transaction_per_chunk
|
|
CREATE UNIQUE INDEX index_test_time_device_idx ON index_test (time) WITH (timescaledb.transaction_per_chunk);
|
|
CREATE UNIQUE INDEX index_test_time_device_idx ON index_test (time, device) WITH(timescaledb.transaction_per_chunk);
|
|
\set ON_ERROR_STOP 1
|
|
|
|
CREATE INDEX index_test_time_device_idx ON index_test (time, device) WITH (timescaledb.transaction_per_chunk);
|
|
|
|
-- Regular index need not cover all partitioning columns
|
|
CREATE INDEX ON index_test (time, temp) WITH (timescaledb.transaction_per_chunk);
|
|
|
|
-- Create another chunk
|
|
INSERT INTO index_test VALUES ('2017-04-20T09:00:01', 1, 17.5);
|
|
|
|
-- New index should have been recursed to chunks
|
|
SELECT * FROM test.show_indexes('index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk') ORDER BY 1,2;
|
|
SELECT * FROM _timescaledb_catalog.chunk_index;
|
|
|
|
ALTER INDEX index_test_time_idx RENAME TO index_test_time_idx2;
|
|
|
|
-- Metadata and index should have changed name
|
|
SELECT * FROM test.show_indexes('index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk') ORDER BY 1,2;
|
|
SELECT * FROM _timescaledb_catalog.chunk_index;
|
|
|
|
DROP INDEX index_test_time_idx2;
|
|
DROP INDEX index_test_time_device_idx;
|
|
|
|
-- Index should have been dropped
|
|
SELECT * FROM test.show_indexes('index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk');
|
|
SELECT * FROM _timescaledb_catalog.chunk_index;
|
|
|
|
-- Create index with long name to see how this is handled on chunks
|
|
CREATE INDEX a_hypertable_index_with_a_very_very_long_name_that_truncates ON index_test (time, temp) WITH (timescaledb.transaction_per_chunk);
|
|
CREATE INDEX a_hypertable_index_with_a_very_very_long_name_that_truncates_2 ON index_test (time, temp) WITH (timescaledb.transaction_per_chunk);
|
|
|
|
SELECT * FROM test.show_indexes('index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk');
|
|
|
|
DROP INDEX a_hypertable_index_with_a_very_very_long_name_that_truncates;
|
|
DROP INDEX a_hypertable_index_with_a_very_very_long_name_that_truncates_2;
|
|
|
|
|
|
SELECT * FROM test.show_indexes('index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk');
|
|
|
|
SELECT * FROM test.show_indexes('index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk');
|
|
|
|
-- Add constraint index
|
|
ALTER TABLE index_test ADD UNIQUE (time, device);
|
|
|
|
SELECT * FROM test.show_indexes('index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk');
|
|
|
|
-- Constraint indexes are added to chunk_index table.
|
|
SELECT * FROM _timescaledb_catalog.chunk_index;
|
|
SELECT * FROM _timescaledb_catalog.chunk_constraint;
|
|
|
|
DROP TABLE index_test;
|
|
|
|
-- Metadata removed
|
|
SELECT * FROM _timescaledb_catalog.chunk_index;
|
|
|
|
-- Test that indexes are planned correctly
|
|
CREATE TABLE index_expr_test(id serial, time timestamptz, temp float, meta int);
|
|
select create_hypertable('index_expr_test', 'time');
|
|
|
|
-- Screw up the attribute numbers
|
|
ALTER TABLE index_expr_test DROP COLUMN id;
|
|
|
|
CREATE INDEX ON index_expr_test (meta) WITH (timescaledb.transaction_per_chunk);
|
|
INSERT INTO index_expr_test VALUES ('2017-01-20T09:00:01', 17.5, 1);
|
|
INSERT INTO index_expr_test VALUES ('2017-01-20T09:00:01', 17.5, 2);
|
|
|
|
|
|
SET enable_seqscan TO false;
|
|
SET enable_bitmapscan TO false;
|
|
EXPLAIN (verbose, costs off)
|
|
SELECT * FROM index_expr_test WHERE meta = 1;
|
|
SELECT * FROM index_expr_test WHERE meta = 1;
|
|
SET enable_seqscan TO default;
|
|
SET enable_bitmapscan TO default;
|
|
|
|
\set ON_ERROR_STOP 0
|
|
-- cannot create a transaction_per_chunk index within a transaction block
|
|
BEGIN;
|
|
CREATE INDEX ON index_expr_test (temp) WITH (timescaledb.transaction_per_chunk);
|
|
ROLLBACK;
|
|
\set ON_ERROR_STOP 1
|
|
|
|
DROP TABLE index_expr_test CASCADE;
|
|
|
|
CREATE TABLE partial_index_test(time INTEGER);
|
|
SELECT create_hypertable('partial_index_test', 'time', chunk_time_interval => 1, create_default_indexes => false);
|
|
|
|
-- create 3 chunks
|
|
INSERT INTO partial_index_test(time) SELECT generate_series(0, 2);
|
|
|
|
select * from partial_index_test order by 1;
|
|
|
|
-- create indexes on only 1 of the chunks
|
|
CREATE INDEX ON partial_index_test (time) WITH (timescaledb.transaction_per_chunk, timescaledb.max_chunks='1');
|
|
|
|
SELECT * FROM test.show_indexes('partial_index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk');
|
|
|
|
-- regerssion test for bug fixed by PR #1008.
|
|
-- this caused an assertion failure when a MergeAppend node contained unsorted children
|
|
SET enable_seqscan TO false;
|
|
SET enable_bitmapscan TO false;
|
|
EXPLAIN (verbose, costs off) SELECT * FROM partial_index_test WHERE time < 2 ORDER BY time LIMIT 2;
|
|
SELECT * FROM partial_index_test WHERE time < 2 ORDER BY time LIMIT 2;
|
|
|
|
-- we can drop the partially created index
|
|
DROP INDEX partial_index_test_time_idx;
|
|
|
|
SELECT * FROM test.show_indexes('partial_index_test');
|
|
SELECT * FROM test.show_indexesp('_timescaledb_internal._hyper%_chunk');
|
|
|
|
EXPLAIN (verbose, costs off) SELECT * FROM partial_index_test WHERE time < 2 ORDER BY time LIMIT 2;
|
|
SELECT * FROM partial_index_test WHERE time < 2 ORDER BY time LIMIT 2;
|
|
|
|
SET enable_seqscan TO true;
|
|
SET enable_bitmapscan TO true;
|