Reputation: 640
I am trying to tune some slow running queries on MariaDB
and have hit a brick wall. I mainly work with Postgres / Oracle and these queries times seem extremely slow to me, but I can't figure out what I'm missing.
The queries in question (see below for examples) are mostly single table queries or with one join / subquery. The table in question has around 19million rows (26.5gb) and has fairly frequent writes.
Table structure and indexes below:
CREATE TABLE products_reviews (
id bigint(20) NOT NULL AUTO_INCREMENT
, product_id int(11) NOT NULL DEFAULT 0
, name varchar(255) NOT NULL DEFAULT ''
, rate tinyint(4) NOT NULL DEFAULT 0
, created_at datetime NOT NULL
, posted_at date NOT NULL
, site varchar(50) NOT NULL
, review text NOT NULL
, details text DEFAULT NULL
, type smallint(6) NOT NULL DEFAULT 0
, notify smallint(6) NOT NULL DEFAULT 1
, notify_daily int(11) NOT NULL DEFAULT 1
, user_id int(11) DEFAULT NULL
, updated_at datetime NOT NULL
, admin_id int(11) DEFAULT NULL
, status smallint(6) NOT NULL DEFAULT 0
, comment text NOT NULL
, notified_at datetime NOT NULL
, reply tinyint(4) NOT NULL DEFAULT 0
, syndicate varchar(100) NOT NULL
, status_user tinyint(4) NOT NULL
, reviewer_id varchar(100) NOT NULL DEFAULT ''
, sentiment text DEFAULT NULL
, sentiment_score float(5,2) NOT NULL DEFAULT 0.00
, ai text DEFAULT NULL
, active tinyint(4) NOT NULL DEFAULT 1
, active_updated_at datetime NOT NULL DEFAULT current_timestamp()
, review_translate text NOT NULL
, price float(6,2) NOT NULL
, tagging text NOT NULL
, site_id int(11) NOT NULL
, hash varchar(255) NOT NULL
, PRIMARY KEY (id)
, KEY created_at (created_at)
, KEY posted_at (posted_at)
, KEY site (site)
, KEY product_id (product_id)
, KEY admin_id (admin_id)
, KEY type (type)
, KEY notify (notify)
, KEY active (active)
, KEY admin_posted (admin_id,posted_at)
, KEY reviewer_id (reviewer_id)
, KEY idx_products_reviews_pspth (product_id,site,posted_at,type,hash)
, KEY idx_products_reviews_pa (product_id,active)
, KEY idx_products_reviews_product_posted (product_id,posted_at)
, KEY idx_products_reviews_product_adminid_postedid (product_id,admin_id,posted_at)
, KEY product_id_2 (product_id,site,active)
) ENGINE=InnoDB AUTO_INCREMENT=27316062 DEFAULT CHARSET=utf8
Mariadb settings below:
alter_algorithm DEFAULT
aria_block_size 8192
aria_checkpoint_interval 30
aria_checkpoint_log_activity 1048576
aria_encrypt_tables OFF
aria_force_start_after_recovery_failures 0
aria_group_commit none
aria_group_commit_interval 0
aria_log_file_size 1073741824
aria_log_purge_type immediate
aria_max_sort_file_size 9223372036853727232
aria_page_checksum ON
aria_pagecache_age_threshold 300
aria_pagecache_buffer_size 134217728
aria_pagecache_division_limit 100
aria_pagecache_file_hash_size 512
aria_recover_options BACKUP,QUICK
aria_repair_threads 1
aria_sort_buffer_size 268434432
aria_stats_method nulls_unequal
aria_sync_log_dir NEWFILE
aria_used_for_temp_tables ON
auto_increment_increment 1
auto_increment_offset 1
autocommit ON
automatic_sp_privileges ON
back_log 100
basedir /usr
big_tables OFF
binlog_annotate_row_events ON
binlog_cache_size 32768
binlog_checksum CRC32
binlog_commit_wait_count 0
binlog_commit_wait_usec 100000
binlog_direct_non_transactional_updates OFF
binlog_file_cache_size 16384
binlog_format MIXED
binlog_optimize_thread_scheduling ON
binlog_row_image FULL
binlog_stmt_cache_size 32768
bulk_insert_buffer_size 8388608
check_constraint_checks ON
column_compression_threshold 100
column_compression_zlib_level 6
column_compression_zlib_strategy DEFAULT_STRATEGY
column_compression_zlib_wrap OFF
completion_type NO_CHAIN
concurrent_insert AUTO
connect_timeout 10
core_file OFF
date_format %Y-%m-%d
datetime_format %Y-%m-%d %H:%i:%s
deadlock_search_depth_long 15
deadlock_search_depth_short 4
deadlock_timeout_long 50000000
deadlock_timeout_short 10000
debug_no_thread_alarm OFF
default_master_connection
default_regex_flags
default_storage_engine InnoDB
default_tmp_storage_engine MyISAM
default_week_format 0
delay_key_write ON
delayed_insert_limit 100
delayed_insert_timeout 300
delayed_queue_size 1000
div_precision_increment 4
encrypt_binlog OFF
encrypt_tmp_disk_tables OFF
encrypt_tmp_files OFF
enforce_storage_engine
eq_range_index_dive_limit 0
error_count 0
event_scheduler OFF
expensive_subquery_limit 100
expire_logs_days 7
explicit_defaults_for_timestamp OFF
external_user
extra_max_connections 1
extra_port 0
flush OFF
flush_time 0
foreign_key_checks ON
ft_boolean_syntax + -><()~*:""&|
ft_max_word_len 84
ft_min_word_len 4
ft_query_expansion_limit 20
ft_stopword_file (built-in)
general_log OFF
group_concat_max_len 1048576
gtid_binlog_pos
gtid_binlog_state
gtid_current_pos
gtid_domain_id 0
gtid_ignore_duplicates OFF
gtid_pos_auto_engines
gtid_seq_no 0
gtid_slave_pos
gtid_strict_mode OFF
have_compress YES
have_crypt YES
have_dynamic_loading YES
have_geometry YES
have_openssl YES
have_profiling YES
have_query_cache YES
have_rtree_keys YES
have_ssl DISABLED
have_symlink YES
histogram_size 0
histogram_type SINGLE_PREC_HB
host_cache_size 633
identity 0
idle_readonly_transaction_timeout 0
idle_transaction_timeout 0
idle_write_transaction_timeout 0
ignore_builtin_innodb OFF
ignore_db_dirs
in_predicate_conversion_threshold 1000
in_transaction 0
init_connect
init_file
init_slave
innodb_adaptive_flushing ON
innodb_adaptive_flushing_lwm 10.000000
innodb_adaptive_hash_index ON
innodb_adaptive_hash_index_parts 8
innodb_adaptive_max_sleep_delay 150000
innodb_autoextend_increment 64
innodb_autoinc_lock_mode 1
innodb_background_scrub_data_check_interval 3600
innodb_background_scrub_data_compressed OFF
innodb_background_scrub_data_interval 604800
innodb_background_scrub_data_uncompressed OFF
innodb_buf_dump_status_frequency 0
innodb_buffer_pool_chunk_size 134217728
innodb_buffer_pool_dump_at_shutdown ON
innodb_buffer_pool_dump_now OFF
innodb_buffer_pool_dump_pct 25
innodb_buffer_pool_filename ib_buffer_pool
innodb_buffer_pool_instances 8
innodb_buffer_pool_load_abort OFF
innodb_buffer_pool_load_at_startup ON
innodb_buffer_pool_load_now OFF
innodb_buffer_pool_size 68719476736
innodb_change_buffer_max_size 25
innodb_change_buffering all
innodb_checksum_algorithm crc32
innodb_checksums ON
innodb_cmp_per_index_enabled OFF
innodb_commit_concurrency 0
innodb_compression_algorithm zlib
innodb_compression_default OFF
innodb_compression_failure_threshold_pct 5
innodb_compression_level 6
innodb_compression_pad_pct_max 50
innodb_concurrency_tickets 5000
innodb_data_file_path ibdata1:12M:autoextend
innodb_data_home_dir
innodb_deadlock_detect ON
innodb_default_encryption_key_id 1
innodb_default_row_format dynamic
innodb_disable_sort_file_cache OFF
innodb_disallow_writes OFF
innodb_doublewrite ON
innodb_encrypt_log OFF
innodb_encrypt_tables OFF
innodb_encrypt_temporary_tables OFF
innodb_encryption_rotate_key_age 1
innodb_encryption_rotation_iops 100
innodb_encryption_threads 0
innodb_fast_shutdown 1
innodb_fatal_semaphore_wait_threshold 600
innodb_file_format
innodb_file_per_table ON
innodb_fill_factor 100
innodb_flush_log_at_timeout 1
innodb_flush_log_at_trx_commit 1
innodb_flush_method fsync
innodb_flush_neighbors 1
innodb_flush_sync ON
innodb_flushing_avg_loops 30
innodb_force_load_corrupted OFF
innodb_force_primary_key OFF
innodb_force_recovery 0
innodb_ft_aux_table
innodb_ft_cache_size 8000000
innodb_ft_enable_diag_print OFF
innodb_ft_enable_stopword ON
innodb_ft_max_token_size 84
innodb_ft_min_token_size 3
innodb_ft_num_word_optimize 2000
innodb_ft_result_cache_limit 2000000000
innodb_ft_server_stopword_table
innodb_ft_sort_pll_degree 2
innodb_ft_total_cache_size 640000000
innodb_ft_user_stopword_table
innodb_idle_flush_pct 100
innodb_immediate_scrub_data_uncompressed OFF
innodb_io_capacity 1000
innodb_io_capacity_max 2000
innodb_large_prefix
innodb_lock_schedule_algorithm fcfs
innodb_lock_wait_timeout 50
innodb_locks_unsafe_for_binlog OFF
innodb_log_buffer_size 16777216
innodb_log_checksums ON
innodb_log_compressed_pages ON
innodb_log_file_size 50331648
innodb_log_files_in_group 2
innodb_log_group_home_dir ./
innodb_log_optimize_ddl ON
innodb_log_write_ahead_size 8192
innodb_lru_scan_depth 1024
innodb_max_dirty_pages_pct 75.000000
innodb_max_dirty_pages_pct_lwm 0.000000
innodb_max_purge_lag 0
innodb_max_purge_lag_delay 0
innodb_max_undo_log_size 10485760
innodb_monitor_disable
innodb_monitor_enable
innodb_monitor_reset
innodb_monitor_reset_all
innodb_old_blocks_pct 37
innodb_old_blocks_time 1000
innodb_online_alter_log_max_size 134217728
innodb_open_files 4096
innodb_optimize_fulltext_only OFF
innodb_page_cleaners 4
innodb_page_size 16384
innodb_prefix_index_cluster_optimization OFF
innodb_print_all_deadlocks OFF
innodb_purge_batch_size 300
innodb_purge_rseg_truncate_frequency 128
innodb_purge_threads 4
innodb_random_read_ahead OFF
innodb_read_ahead_threshold 56
innodb_read_io_threads 4
innodb_read_only OFF
innodb_replication_delay 0
innodb_rollback_on_timeout OFF
innodb_rollback_segments 128
innodb_scrub_log OFF
innodb_scrub_log_speed 256
innodb_sort_buffer_size 1048576
innodb_spin_wait_delay 4
innodb_stats_auto_recalc ON
innodb_stats_include_delete_marked OFF
innodb_stats_method nulls_equal
innodb_stats_modified_counter 0
innodb_stats_on_metadata OFF
innodb_stats_persistent ON
innodb_stats_persistent_sample_pages 20
innodb_stats_sample_pages 8
innodb_stats_traditional ON
innodb_stats_transient_sample_pages 8
innodb_status_output OFF
innodb_status_output_locks OFF
innodb_strict_mode OFF
innodb_sync_array_size 1
innodb_sync_spin_loops 30
innodb_table_locks ON
innodb_temp_data_file_path ibtmp1:12M:autoextend
innodb_thread_concurrency 0
innodb_thread_sleep_delay 10000
innodb_tmpdir
innodb_undo_directory ./
innodb_undo_log_truncate OFF
innodb_undo_logs 128
innodb_undo_tablespaces 0
innodb_use_atomic_writes ON
innodb_use_native_aio ON
innodb_version 10.3.21
innodb_write_io_threads 4
insert_id 0
interactive_timeout 3600
join_buffer_size 1048576
join_buffer_space_limit 2097152
join_cache_level 2
keep_files_on_create OFF
key_buffer_size 67108864
key_cache_age_threshold 300
key_cache_block_size 1024
key_cache_division_limit 100
key_cache_file_hash_size 512
key_cache_segments 0
large_files_support ON
large_page_size 0
large_pages OFF
last_gtid
last_insert_id 0
lc_messages en_US
lc_messages_dir
lc_time_names en_US
license GPL
local_infile ON
lock_wait_timeout 86400
locked_in_memory OFF
log_bin OFF
log_bin_basename
log_bin_compress OFF
log_bin_compress_min_len 256
log_bin_index
log_bin_trust_function_creators OFF
log_disabled_statements sp
log_error /var/log/mariadb/mariadb.log
log_output FILE
log_queries_not_using_indexes OFF
log_slave_updates OFF
log_slow_admin_statements ON
log_slow_disabled_statements sp
log_slow_filter admin,filesort,filesort_on_disk,filesort_priority_queue,full_join,full_scan,query_cache,query_cache_miss,tmp_table,tmp_table_on_disk
log_slow_rate_limit 1
log_slow_slave_statements ON
log_slow_verbosity
log_tc_size 24576
log_warnings 2
long_query_time 1.000000
low_priority_updates OFF
lower_case_file_system OFF
lower_case_table_names 0
master_verify_checksum OFF
max_allowed_packet 67108864
max_binlog_cache_size 18446744073709547520
max_binlog_size 1073741824
max_binlog_stmt_cache_size 18446744073709547520
max_connect_errors 10000
max_connections 600
max_delayed_threads 20
max_digest_length 1024
max_error_count 64
max_heap_table_size 67108864
max_insert_delayed_threads 20
max_join_size 18446744073709551615
max_length_for_sort_data 1024
max_long_data_size 67108864
max_prepared_stmt_count 16382
max_recursive_iterations 4294967295
max_relay_log_size 1073741824
max_seeks_for_key 4294967295
max_session_mem_used 9223372036854775807
max_sort_length 1024
max_sp_recursion_depth 0
max_statement_time 0.000000
max_tmp_tables 32
max_user_connections 0
max_write_lock_count 4294967295
metadata_locks_cache_size 1024
metadata_locks_hash_instances 8
min_examined_row_limit 0
mrr_buffer_size 262144
multi_range_count 256
myisam_block_size 1024
myisam_data_pointer_size 6
myisam_max_sort_file_size 9223372036853727232
myisam_mmap_size 18446744073709551615
myisam_recover_options BACKUP,QUICK
myisam_repair_threads 1
myisam_sort_buffer_size 134217728
myisam_stats_method NULLS_UNEQUAL
myisam_use_mmap OFF
mysql56_temporal_format ON
net_buffer_length 16384
net_read_timeout 30
net_retry_count 10
net_write_timeout 60
open_files_limit 65535
optimizer_prune_level 1
optimizer_search_depth 62
optimizer_selectivity_sampling_limit 100
optimizer_switch index_merge=off,index_merge_union=on,index_merge_sort_union=on,index_merge_intersection=on,index_merge_sort_intersection=off,engine_condition_pushdown=off,index_condition_pushdown=on,derived_merge=on,derived_with_keys=on,firstmatch=on,loosescan=on,materialization=on,in_to_exists=on,semijoin=on,partial_match_rowid_merge=on,partial_match_table_scan=on,subquery_cache=on,mrr=off,mrr_cost_based=off,mrr_sort_keys=off,outer_join_with_cache=on,semijoin_with_cache=on,join_cache_incremental=on,join_cache_hashed=on,join_cache_bka=on,optimize_join_buffer_size=off,table_elimination=on,extended_keys=on,exists_to_in=on,orderby_uses_equalities=on,condition_pushdown_for_derived=on,split_materialized=on
optimizer_use_condition_selectivity 1
performance_schema OFF
performance_schema_accounts_size -1
performance_schema_digests_size -1
performance_schema_events_stages_history_long_size -1
performance_schema_events_stages_history_size -1
performance_schema_events_statements_history_long_size -1
performance_schema_events_statements_history_size -1
performance_schema_events_waits_history_long_size -1
performance_schema_events_waits_history_size -1
performance_schema_hosts_size -1
performance_schema_max_cond_classes 80
performance_schema_max_cond_instances -1
performance_schema_max_digest_length 1024
performance_schema_max_file_classes 50
performance_schema_max_file_handles 32768
performance_schema_max_file_instances -1
performance_schema_max_mutex_classes 200
performance_schema_max_mutex_instances -1
performance_schema_max_rwlock_classes 40
performance_schema_max_rwlock_instances -1
performance_schema_max_socket_classes 10
performance_schema_max_socket_instances -1
performance_schema_max_stage_classes 160
performance_schema_max_statement_classes 200
performance_schema_max_table_handles -1
performance_schema_max_table_instances -1
performance_schema_max_thread_classes 50
performance_schema_max_thread_instances -1
performance_schema_session_connect_attrs_size -1
performance_schema_setup_actors_size 100
performance_schema_setup_objects_size 100
performance_schema_users_size -1
pid_file /var/run/mariadb/mariadb.pid
plugin_dir /usr/lib64/mysql/plugin/
plugin_maturity gamma
port 3306
preload_buffer_size 32768
profiling OFF
profiling_history_size 15
progress_report_time 5
protocol_version 10
proxy_protocol_networks
proxy_user
pseudo_slave_mode OFF
pseudo_thread_id 1852681
query_alloc_block_size 16384
query_cache_limit 1048576
query_cache_min_res_unit 4096
query_cache_size 1048576
query_cache_strip_comments OFF
query_cache_type OFF
query_cache_wlock_invalidate OFF
query_prealloc_size 24576
rand_seed1 93838501
rand_seed2 36662033
range_alloc_block_size 4096
read_binlog_speed_limit 0
read_buffer_size 1048576
read_only OFF
read_rnd_buffer_size 1048576
relay_log
relay_log_basename
relay_log_index
relay_log_info_file relay-log.info
relay_log_purge ON
relay_log_recovery OFF
relay_log_space_limit 17179869184
replicate_annotate_row_events ON
replicate_do_db
replicate_do_table
replicate_events_marked_for_skip REPLICATE
replicate_ignore_db
replicate_ignore_table
replicate_wild_do_table
replicate_wild_ignore_table
report_host
report_password
report_port 3306
report_user
rowid_merge_buff_size 8388608
secure_auth ON
secure_file_priv
secure_timestamp NO
server_id 1
session_track_schema ON
session_track_state_change OFF
session_track_system_variables autocommit,character_set_client,character_set_connection,character_set_results,time_zone
session_track_transaction_info OFF
skip_external_locking ON
skip_name_resolve ON
skip_networking OFF
skip_parallel_replication OFF
skip_replication OFF
skip_show_database OFF
slave_compressed_protocol OFF
slave_ddl_exec_mode IDEMPOTENT
slave_domain_parallel_threads 0
slave_exec_mode STRICT
slave_load_tmpdir /san/svc-fs/mysqltmp
slave_max_allowed_packet 1073741824
slave_net_timeout 60
slave_parallel_max_queued 131072
slave_parallel_mode conservative
slave_parallel_threads 0
slave_parallel_workers 0
slave_run_triggers_for_rbr NO
slave_skip_errors OFF
slave_sql_verify_checksum ON
slave_transaction_retries 10
slave_transaction_retry_errors 1213,1205
slave_transaction_retry_interval 0
slave_type_conversions
slow_launch_time 2
slow_query_log ON
slow_query_log_file /san/svc-fs/mysqllogs/slow-log
socket /san/svc-fs/mysql/mysql.sock
sort_buffer_size 1048576
sql_auto_is_null OFF
sql_big_selects ON
sql_buffer_result OFF
sql_log_bin ON
sql_log_off OFF
sql_mode NO_ENGINE_SUBSTITUTION
sql_notes ON
sql_quote_show_create ON
sql_safe_updates OFF
sql_select_limit 18446744073709551615
sql_slave_skip_counter 0
sql_warnings OFF
ssl_ca
ssl_capath
ssl_cert
ssl_cipher
ssl_crl
ssl_crlpath
ssl_key
standard_compliant_cte ON
storage_engine InnoDB
stored_program_cache 256
strict_password_validation ON
sync_binlog 0
sync_frm ON
sync_master_info 10000
sync_relay_log 10000
sync_relay_log_info 10000
system_time_zone CST
system_versioning_alter_history ERROR
system_versioning_asof DEFAULT
table_definition_cache 4096
table_open_cache 4096
table_open_cache_instances 8
tcp_keepalive_interval 0
tcp_keepalive_probes 0
tcp_keepalive_time 0
thread_cache_size 16
thread_concurrency 10
thread_handling one-thread-per-connection
thread_pool_idle_timeout 60
thread_pool_max_threads 65536
thread_pool_oversubscribe 3
thread_pool_prio_kickup_timer 1000
thread_pool_priority auto
thread_pool_size 16
thread_pool_stall_limit 500
thread_stack 299008
time_format %H:%i:%s
time_zone SYSTEM
timed_mutexes OFF
timestamp 1644789513.087889
tmp_disk_table_size 18446744073709551615
tmp_memory_table_size 33554432
tmp_table_size 33554432
tmpdir /san/svc-fs/mysqltmp
transaction_alloc_block_size 8192
transaction_prealloc_size 4096
tx_isolation REPEATABLE-READ
tx_read_only OFF
unique_checks ON
updatable_views_with_limit YES
use_stat_tables NEVER
userstat OFF
version 10.3.21-MariaDB-log
version_comment MariaDB Server
version_compile_machine x86_64
version_compile_os Linux
version_malloc_library jemalloc 3.6.0-0-g46c0af68bd248b04df75e4f92d5fb804c3d75340
version_source_revision e5e5877740f248de848219ee3a1d2881cd5c5b82
version_ssl_library OpenSSL 1.0.2k-fips 26 Jan 2017
wait_timeout 600
warning_count 0
wsrep_osu_method TOI
wsrep_auto_increment_control ON
wsrep_causal_reads OFF
wsrep_certification_rules strict
wsrep_certify_nonpk ON
wsrep_cluster_address
wsrep_cluster_name my_wsrep_cluster
wsrep_convert_lock_to_trx OFF
wsrep_data_home_dir /san/svc-fs/mysql/
wsrep_dbug_option
wsrep_debug OFF
wsrep_desync OFF
wsrep_dirty_reads OFF
wsrep_drupal_282555_workaround OFF
wsrep_forced_binlog_format NONE
wsrep_gtid_domain_id 0
wsrep_gtid_mode OFF
wsrep_load_data_splitting ON
wsrep_log_conflicts OFF
wsrep_max_ws_rows 0
wsrep_max_ws_size 2147483647
wsrep_mysql_replication_bundle 0
wsrep_node_address
wsrep_node_incoming_address AUTO
wsrep_notify_cmd
wsrep_on OFF
wsrep_patch_version wsrep_25.24
wsrep_provider none
wsrep_provider_options
wsrep_recover OFF
wsrep_reject_queries NONE
wsrep_replicate_myisam OFF
wsrep_restart_slave OFF
wsrep_retry_autocommit 1
wsrep_slave_fk_checks ON
wsrep_slave_uk_checks OFF
wsrep_slave_threads 1
wsrep_sst_auth
wsrep_sst_donor
wsrep_sst_donor_rejects_queries OFF
wsrep_sst_method rsync
wsrep_sst_receive_address AUTO
wsrep_start_position 00000000-0000-0000-0000-000000000000:-1
wsrep_sync_wait 0
The queries filters can be selected on the front end by a user so the where clause can vary a bit. They will all contain a product_id at least. The speed to read from disk seems absurdly slow for the number of rows being fetched.
Example queries: This one takes 400seconds to return
Edit - Below query is in an incorrect format which must have happened when creating the post. Original is below
SELECT SQL_NO_CACHE
pr.rate as star, pr.product_id, pr.site, COUNT(pr.id) AS count
FROM products_reviews pr
JOIN products p ON p.id = pr.product_id
AND pr.active > 0
AND pr.posted_at >= '2020-10-26'
AND pr.posted_at <= '2022-01-24 23:59'
AND pr.admin_id = 3598
GROUP BY pr.rate, pr.product_id, pr.site;
SELECT rate as star, product_id, site, COUNT(id) AS count
FROM products_reviews
WHERE product_id IN (SELECT id FROM products WHERE admin_id = 3598)
AND active >0 AND posted_at >= '2020-10-26'
AND posted_at <= '2022-01-24 23:59'
AND admin_id = 3598
GROUP BY rate, product_id, site
Explain plan
{
"query_block": {
"select_id": 1,
"filesort": {
"sort_key": "products_reviews.rate, products_reviews.product_id, products_reviews.site",
"temporary_table": {
"table": {
"table_name": "products_reviews",
"access_type": "range",
"possible_keys": [
"posted_at",
"product_id",
"admin_id",
"active",
"admin_posted",
"idx_products_reviews_pspth",
"idx_products_reviews_pa",
"idx_products_reviews_product_posted",
"idx_products_reviews_product_adminid_postedid",
"product_id_2",
"idx_products_aaprps_covering",
"idx_products_aprps_covering_t"
],
"key": "idx_products_aaprps_covering",
"key_length": "6",
"used_key_parts": ["admin_id", "active"],
"rows": 1826162,
"filtered": 100,
"attached_condition": "products_reviews.admin_id = 3598 and products_reviews.active > 0 and products_reviews.posted_at >= '2020-10-26' and products_reviews.posted_at <= '2022-01-24 23:59'",
"using_index": true
},
"table": {
"table_name": "products",
"access_type": "eq_ref",
"possible_keys": ["PRIMARY", "id", "admin_id"],
"key": "PRIMARY",
"key_length": "8",
"used_key_parts": ["id"],
"ref": ["metabase.products_reviews.product_id"],
"rows": 1,
"filtered": 100,
"attached_condition": "products.admin_id = 3598 and products_reviews.product_id = products.`id`"
}
}
}
}
}
I ran profiling on the transaction and the whole time is spent on "Sending Data".
Here is a second query, this one takes around 32 seconds.
SELECT site, MIN(posted_at) as posted_at, AVG(IF(rate>0, rate, NULL)) as rate, COUNT(IF(type=0, 1, NULL)) as count, COUNT(IF(type=0 AND syndicate='', 1, NULL)) as count_ns, AVG(IF(rate>0 AND syndicate='', rate, NULL)) as rate_ns, COUNT(IF(type=1, 1, NULL)) as count_qe
FROM products_reviews
WHERE product_id = '24828'
AND active >0
GROUP BY site
Explain Plan
{
"query_block": {
"select_id": 1,
"table": {
"table_name": "products_reviews",
"access_type": "ref",
"possible_keys": [
"product_id",
"active",
"idx_products_reviews_pspth",
"idx_products_reviews_pa",
"idx_products_reviews_product_posted",
"idx_products_reviews_product_adminid_postedid",
"product_id_2"
],
"key": "product_id_2",
"key_length": "4",
"used_key_parts": ["product_id"],
"ref": ["const"],
"rows": 6575,
"filtered": 100,
"attached_condition": "products_reviews.product_id <=> '24828' and products_reviews.active > 0"
}
}
{
"query_block": {
"select_id": 1,
"table": {
"table_name": "products_reviews",
"access_type": "ref",
"possible_keys": [
"product_id",
"active",
"idx_products_reviews_pspth",
"idx_products_reviews_pa",
"idx_products_reviews_product_posted",
"idx_products_reviews_product_adminid_postedid",
"product_id_2"
],
"key": "product_id_2",
"key_length": "4",
"used_key_parts": ["product_id"],
"ref": ["const"],
"rows": 6575,
"filtered": 100,
"attached_condition": "products_reviews.product_id <=> '24828' and products_reviews.active > 0"
}
}
}
Possibly partially related to cardinality. Here's a count of rows, distinct product_id
and distinct posted_at
.
count(*) 19128090
count(distinct(product_id)) 28821
count(distinct(posted_at)) 7551
The server has 96gb RAM and a 480GB SATA SSD. The buffer pool settings look well optimized to me. I tried increasing innodb_io_capacity
but that had a negative impact.
If there is nothing I can do to speed up the queries with regards query tuning or parameter tuning, I wondered about keeping the table in memory at all times. They seem to get flushed at least once a day (probably around the time backups are taken). Would it be sensible to have a job that runs once every couple of hours to select * from the table?
Many thanks.
Upvotes: 1
Views: 1833
Reputation: 142298
`products is only sort of used. If you don't really need it, get rid of it. I see two things that can matter:
products
the JOIN
will avoid including it.products
, then the COUNT
will increase.Replace
JOIN products p ON p.id = pr.product_id
AND
with
WHERE
As a rule:
ON
to specify how the tables are related.WHERE
to specify which rows to keep.Another INDEX tip: When you need INDEX(a,b)
, you can do without INDEX(a)
. I am thinking of (admin_id, posted_at)
If active
is only "0" or "1", then use AND active = 1
, plus change to INDEX(admin_id, active, posted_at)
"indexes could only contain one equality operator" -- The rule is more like "Indexes should contain only one range operator, and it needs to be last". (Note: "covering" is a deliberate exception.)
(I agree that the 'covering' index may help, but I generally avoid indexes with lots of columns.)
I have the master of that MariaDB link (that RLOG mentioned); it has been updated a few times here: http://mysql.rjweb.org/doc.php/index_cookbook_mysql
Upvotes: 1
Reputation: 108676
It's a good question you ask. My answer has a few dimensions.
Server provisioning: Your machine is adequate, and possibly overprovisioned. You have allocated 2/3 of your server's RAM to your 64GiB InnoDB buffer pool (innodb_buffer_pool_size) and that's probably enough to hold your tables. You could allocate another 8 - 12 GiB to it and still have RAM left over for your server's other functions. But buffer pool size is almost certainly not your issue.
Many indexes: You have quite a few single-column indexes. Are you sure they are all needed? Generally speaking, we create indexes to support particular query patterns, not because they might be helpful someday. If you can drop some of your indexes, do so. And, indexes slow down INSERTs and UPDATEs because both the table data and indexes are affected. Which brings us to ...
Frequent updates: It's possible that your update operations slow down the reporting-style queries you showed us in your question because the server is struggling to give consistent results (using locks) as updates happen. Try giving this SQL statement right before your too-slow query.
SET TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
This causes a "dirty read". It cuts down on locking, but gives you the chance you'll get an older version of rows that are being actively updated. That's usually no problem for reporting-style queries. Try it and see how much it helps.
Equality and range filters, covering BTREE indexes. Consider your first query, repeated here:
SELECT SQL_NO_CACHE
pr.rate as star, pr.product_id, pr.site, COUNT(pr.id) AS count
FROM products_reviews pr
JOIN products p ON p.id = pr.product_id
AND pr.active > 0
AND pr.posted_at >= '2020-10-26'
AND pr.posted_at <= '2022-01-24 23:59'
AND pr.admin_id = 3598
GROUP BY pr.rate, pr.product_id, pr.site;
You filter product_reviews
on admin_id
for equality, then active
and posted at
for ranges, then you look up rows in products
, then group by rate
, product_id
, and site
.
Try using this so-called covering index.
CREATE INDEX rate_product_site_rollup ON products_review
(admin_id, posted_at, active, rate, product_id, site);
It contains the data your query needs from the products_review
table. To satisfy the query, the database can random-access the index to the first eligible row (admin_id
= nnnn and the first eligible posted_at
row. It can then scan the index sequentially to filter out active
<= 0 rows. Finally, your GROUP BY columns are there already sorted in an order suitable for quick grouping.
You might consider a similar covering index for your second query.
(Some makes and versions of database servers allow INCLUDE(col, col, col)
clauses in index declarations, but not MariaDB / MySQL. INCLUDE
is very useful for covering indexes, but including the columns in the index works too.)
More than one range filter: BTREE indexes do their fastest work when you only have one range filter in your query. If your business rules allowed you to filter on pr.active = 1
then this index would satisfy your query even faster.
CREATE INDEX rate_product_site_rollup ON products_review
(admin_id, active, posted_at, rate, product_id, site);
Extraneous JOIN: You JOIN to your products
table, but don't use any columns from it. So the JOIN serves only one purpose: to exclude reviews that have no matching product. Is this step necessary?
Note: pr.posted_at < '2022-01-24' + INTERVAL 1 DAY
, with <
, is a bit more reliable on the almost-midnight edge case than pr.posted_at <= '2022-01-24 23:59'
Upvotes: 1