mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-12-02 12:32:04 +00:00
80 lines
3.9 KiB
Bash
Executable File
80 lines
3.9 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
|
|
# shellcheck disable=SC2154
|
|
|
|
unset CLICKHOUSE_LOG_COMMENT
|
|
|
|
CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
|
|
# shellcheck source=../shell_config.sh
|
|
. "$CURDIR"/../shell_config.sh
|
|
|
|
|
|
check_replicas_read_in_order() {
|
|
# to check this we actually look for at least one log message from MergeTreeInOrderSelectProcessor.
|
|
# hopefully logger's names are a bit more stable than log messages itself
|
|
#
|
|
# NOTE: lack of "current_database = '$CLICKHOUSE_DATABASE'" filter is made on purpose
|
|
$CLICKHOUSE_CLIENT -nq "
|
|
SYSTEM FLUSH LOGS;
|
|
|
|
SELECT COUNT() > 0
|
|
FROM system.text_log
|
|
WHERE query_id IN (SELECT query_id FROM system.query_log WHERE query_id != '$1' AND initial_query_id = '$1' AND event_date >= yesterday())
|
|
AND event_date >= yesterday() AND logger_name = 'MergeTreeInOrderSelectProcessor'"
|
|
}
|
|
|
|
# replicas should use reading in order following initiator's decision to execute aggregation in order.
|
|
# at some point we had a bug in this logic (see https://github.com/ClickHouse/ClickHouse/pull/45892#issue-1566140414)
|
|
test1() {
|
|
query_id="query_id_memory_bound_merging_$RANDOM$RANDOM"
|
|
$CLICKHOUSE_CLIENT --query_id="$query_id" -nq "
|
|
SET cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
|
|
|
|
SELECT URL, EventDate, max(URL)
|
|
FROM remote(test_cluster_one_shard_two_replicas, test.hits)
|
|
WHERE CounterID = 1704509 AND UserID = 4322253409885123546
|
|
GROUP BY CounterID, URL, EventDate
|
|
ORDER BY URL, EventDate
|
|
LIMIT 5 OFFSET 10
|
|
SETTINGS optimize_aggregation_in_order = 1, enable_memory_bound_merging_of_aggregation_results = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, max_parallel_replicas = 3, use_hedged_requests = 0"
|
|
check_replicas_read_in_order $query_id
|
|
}
|
|
|
|
# replicas should use reading in order following initiator's decision to execute aggregation in order.
|
|
# at some point we had a bug in this logic (see https://github.com/ClickHouse/ClickHouse/pull/45892#issue-1566140414)
|
|
test2() {
|
|
query_id="query_id_memory_bound_merging_$RANDOM$RANDOM"
|
|
$CLICKHOUSE_CLIENT --query_id="$query_id" -nq "
|
|
SET cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
|
|
|
|
SELECT URL, EventDate, max(URL)
|
|
FROM remote(test_cluster_one_shard_two_replicas, test.hits)
|
|
WHERE CounterID = 1704509 AND UserID = 4322253409885123546
|
|
GROUP BY URL, EventDate
|
|
ORDER BY URL, EventDate
|
|
LIMIT 5 OFFSET 10
|
|
SETTINGS optimize_aggregation_in_order = 1, enable_memory_bound_merging_of_aggregation_results = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, max_parallel_replicas = 3, use_hedged_requests = 0, query_plan_aggregation_in_order = 1"
|
|
check_replicas_read_in_order $query_id
|
|
}
|
|
|
|
test3() {
|
|
$CLICKHOUSE_CLIENT -nq "
|
|
SET cluster_for_parallel_replicas = 'test_cluster_one_shard_three_replicas_localhost';
|
|
SET max_threads = 16, prefer_localhost_replica = 1, read_in_order_two_level_merge_threshold = 1000, query_plan_aggregation_in_order = 1, distributed_aggregation_memory_efficient = 1;
|
|
|
|
SELECT replaceRegexpOne(explain, '^ *(\w+).*', '\\1')
|
|
FROM (
|
|
EXPLAIN PIPELINE
|
|
SELECT URL, EventDate, max(URL)
|
|
FROM test.hits
|
|
WHERE CounterID = 1704509 AND UserID = 4322253409885123546
|
|
GROUP BY URL, EventDate
|
|
SETTINGS optimize_aggregation_in_order = 1, enable_memory_bound_merging_of_aggregation_results = 1, allow_experimental_parallel_reading_from_replicas = 1, parallel_replicas_for_non_replicated_merge_tree = 1, max_parallel_replicas = 3, use_hedged_requests = 0
|
|
)
|
|
WHERE explain LIKE '%Aggr%Transform%' OR explain LIKE '%InOrder%'"
|
|
}
|
|
|
|
test1
|
|
test2
|
|
test3
|