Inserts gradually slowing down

Hi,

I’m designing a database for storing results of some experiment. The database has only 5 tables and currently I’m working with only the one for storing experiment data, trying to import 1.4GB of numerical data from text files. The table looks as follows:

‘CREATE TABLE t_data (
a int(10) unsigned NOT NULL,
b tinyint(3) unsigned NOT NULL,
c int(10) unsigned NOT NULL,
double NOT NULL,
int(10) unsigned NOT NULL,
int(10) unsigned NOT NULL,
tinyint(1) NOT NULL,
float NOT NULL,
float NOT NULL,
float NOT NULL,
float NOT NULL,
int(10) unsigned NOT NULL,
int(10) unsigned NOT NULL,
int(10) unsigned NOT NULL,
smallint(5) unsigned NOT NULL,
float NOT NULL,
float NOT NULL,
float NOT NULL,
float NOT NULL,
float NOT NULL,
tinyint(1) NOT NULL,
PRIMARY KEY USING BTREE (a,b,c)
KEY fk_data_a (a),
CONSTRAINT fk_data_a FOREIGN KEY (a) REFERENCES t_experiments (a)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_bin’

I import the data by parsing 1000 rows of text and inserting them at once with multiple VALUES. The problem is that after some amount of data the speed decreases considerably. At the beginning it is about 2000-2500KB/s and after importing 1GB it drops to something like 300KB/s. The total amount of data may reach 300 or even 500GB so I’m afraid what the performance would be like if after 1GB it’s already so low.
Currently, I’m testing it on my laptop which has a Penryn processor and 3GB of RAM, but the machine on which it has to run is some Celeron with 512MB of RAM.

What can be done to improve the performance?

mysql> show global variables;
±--------------------------------±------------------------ ----+
| Variable_name | Value |
±--------------------------------±------------------------ ----+
| auto_increment_increment | 1 |
| auto_increment_offset | 1 |
| automatic_sp_privileges | ON |
| back_log | 50 |
| basedir | /usr/ |
| binlog_cache_size | 32768 |
| bulk_insert_buffer_size | 8388608 |
| character_set_client | latin1 |
| character_set_connection | latin1 |
| character_set_database | latin1 |
| character_set_filesystem | binary |
| character_set_results | latin1 |
| character_set_server | latin1 |
| character_set_system | utf8 |
| character_sets_dir | /usr/share/mysql/charsets/ |
| collation_connection | latin1_swedish_ci |
| collation_database | latin1_swedish_ci |
| collation_server | latin1_swedish_ci |
| completion_type | 0 |
| concurrent_insert | 1 |
| connect_timeout | 5 |
| datadir | /var/lib/mysql/ |
| date_format | %Y-%m-%d |
| datetime_format | %Y-%m-%d %H:%i:%s |
| default_week_format | 0 |
| delay_key_write | ON |
| delayed_insert_limit | 100 |
| delayed_insert_timeout | 300 |
| delayed_queue_size | 1000 |
| div_precision_increment | 4 |
| keep_files_on_create | OFF |
| engine_condition_pushdown | OFF |
| expire_logs_days | 10 |
| flush | OFF |
| flush_time | 0 |
| ft_boolean_syntax | + -><()~*:“”&| |
| ft_max_word_len | 84 |
| ft_min_word_len | 4 |
| ft_query_expansion_limit | 20 |
| ft_stopword_file | (built-in) |
| group_concat_max_len | 1024 |
| have_archive | YES |
| have_bdb | NO |
| have_blackhole_engine | YES |
| have_compress | YES |
| have_crypt | YES |
| have_csv | YES |
| have_dynamic_loading | YES |
| have_example_engine | NO |
| have_federated_engine | YES |
| have_geometry | YES |
| have_innodb | DISABLED |
| have_isam | NO |
| have_merge_engine | YES |
| have_ndbcluster | DISABLED |
| have_openssl | DISABLED |
| have_ssl | DISABLED |
| have_query_cache | YES |
| have_raid | NO |
| have_rtree_keys | YES |
| have_symlink | YES |
| hostname | julek-laptop |
| init_connect | |
| init_file | |
| init_slave | |
| innodb_additional_mem_pool_size | 1048576 |
| innodb_autoextend_increment | 8 |
| innodb_buffer_pool_awe_mem_mb | 0 |
| innodb_buffer_pool_size | 268435456 |
| innodb_checksums | ON |
| innodb_commit_concurrency | 0 |
| innodb_concurrency_tickets | 500 |
| innodb_data_file_path | ibdata1:10M:autoextend |
| innodb_data_home_dir | |
| innodb_doublewrite | ON |
| innodb_fast_shutdown | 1 |
| innodb_file_io_threads | 4 |
| innodb_file_per_table | ON |
| innodb_flush_log_at_trx_commit | 2 |
| innodb_flush_method | O_DIRECT |
| innodb_force_recovery | 0 |
| innodb_lock_wait_timeout | 50 |
| innodb_locks_unsafe_for_binlog | OFF |
| innodb_log_arch_dir | |
| innodb_log_archive | OFF |
| innodb_log_buffer_size | 4194304 |
| innodb_log_file_size | 134217728 |
| innodb_log_files_in_group | 2 |
| innodb_log_group_home_dir | ./ |
| innodb_max_dirty_pages_pct | 90 |
| innodb_max_purge_lag | 0 |
| innodb_mirrored_log_groups | 1 |
| innodb_open_files | 300 |
| innodb_rollback_on_timeout | OFF |
| innodb_support_xa | ON |
| innodb_sync_spin_loops | 20 |
| innodb_table_locks | ON |
| innodb_thread_concurrency | 8 |
| innodb_thread_sleep_delay | 10000 |
| interactive_timeout | 28800 |
| join_buffer_size | 131072 |
| key_buffer_size | 268435456 |
| key_cache_age_threshold | 300 |
| key_cache_block_size | 1024 |
| key_cache_division_limit | 100 |
| language | /usr/share/mysql/english/ |
| large_files_support | ON |
| large_page_size | 0 |
| large_pages | OFF |
| lc_time_names | en_US |
| license | GPL |
| local_infile | ON |
| locked_in_memory | OFF |
| log | OFF |
| log_bin | OFF |
| log_bin_trust_function_creators | OFF |
| log_error | |
| log_queries_not_using_indexes | OFF |
| log_slave_updates | OFF |
| log_slow_queries | OFF |
| log_warnings | 1 |
| long_query_time | 10 |
| low_priority_updates | OFF |
| lower_case_file_system | OFF |
| lower_case_table_names | 0 |
| max_allowed_packet | 16776192 |
| max_binlog_cache_size | 4294967295 |
| max_binlog_size | 104857600 |
| max_connect_errors | 10 |
| max_connections | 100 |
| max_delayed_threads | 20 |
| max_error_count | 64 |
| max_heap_table_size | 16777216 |
| max_insert_delayed_threads | 20 |
| max_join_size | 18446744073709551615 |
| max_length_for_sort_data | 1024 |
| max_prepared_stmt_count | 16382 |
| max_relay_log_size | 0 |
| max_seeks_for_key | 4294967295 |
| max_sort_length | 1024 |
| max_sp_recursion_depth | 0 |
| max_tmp_tables | 32 |
| max_user_connections | 0 |
| max_write_lock_count | 4294967295 |
| multi_range_count | 256 |
| myisam_data_pointer_size | 6 |
| myisam_max_sort_file_size | 2147483647 |
| myisam_recover_options | OFF |
| myisam_repair_threads | 1 |
| myisam_sort_buffer_size | 8388608 |
| myisam_stats_method | nulls_unequal |
| ndb_autoincrement_prefetch_sz | 32 |
| ndb_force_send | ON |
| ndb_use_exact_count | ON |
| ndb_use_transactions | ON |
| ndb_cache_check_time | 0 |
| ndb_connectstring | |
| net_buffer_length | 16384 |
| net_read_timeout | 30 |
| net_retry_count | 10 |
| net_write_timeout | 60 |
| new | OFF |
| old_passwords | ON |
| open_files_limit | 1024 |
| optimizer_prune_level | 1 |
| optimizer_search_depth | 62 |
| pid_file | /var/run/mysqld/mysqld.pid |
| port | 3304 |
| preload_buffer_size | 32768 |
| profiling | OFF |
| profiling_history_size | 15 |
| protocol_version | 10 |
| query_alloc_block_size | 8192 |
| query_cache_limit | 1048576 |
| query_cache_min_res_unit | 4096 |
| query_cache_size | 16777216 |
| query_cache_type | ON |
| query_cache_wlock_invalidate | OFF |
| query_prealloc_size | 8192 |
| range_alloc_block_size | 2048 |
| read_buffer_size | 131072 |
| read_only | OFF |
| read_rnd_buffer_size | 262144 |
| relay_log_purge | ON |
| relay_log_space_limit | 0 |
| rpl_recovery_rank | 0 |
| secure_auth | OFF |
| secure_file_priv | |
| server_id | 0 |
| skip_external_locking | ON |
| skip_networking | OFF |
| skip_show_database | OFF |
| slave_compressed_protocol | OFF |
| slave_load_tmpdir | /tmp/ |
| slave_net_timeout | 3600 |
| slave_skip_errors | OFF |
| slave_transaction_retries | 10 |
| slow_launch_time | 2 |
| socket | /var/run/mysqld/mysqld.sock |
| sort_buffer_size | 2097144 |
| sql_big_selects | ON |
| sql_mode | |
| sql_notes | ON |
| sql_warnings | OFF |
| ssl_ca | |
| ssl_capath | |
| ssl_cert | |
| ssl_cipher | |
| ssl_key | |
| storage_engine | MyISAM |
| sync_binlog | 0 |
| sync_frm | ON |
| system_time_zone | IST |
| table_cache | 64 |
| table_lock_wait_timeout | 50 |
| table_type | MyISAM |
| thread_cache_size | 8 |
| thread_stack | 131072 |
| time_format | %H:%i:%s |
| time_zone | SYSTEM |
| timed_mutexes | OFF |
| tmp_table_size | 33554432 |
| tmpdir | /tmp |
| transaction_alloc_block_size | 8192 |
| transaction_prealloc_size | 4096 |
| tx_isolation | READ-COMMITTED |
| updatable_views_with_limit | YES |
| version | 5.0.51a-3ubuntu5.2 |
| version_comment | (Ubuntu) |
| version_compile_machine | i486 |
| version_compile_os | debian-linux-gnu |
| wait_timeout | 28800 |
±--------------------------------±------------------------ ----+

Regards,
Juliusz Gonera

I think the most obvious thing would be to disable your indexes during the import, and rebuild them when you’re done. Also, if you don’t need the transactional support of InnoDB, you could make your table MyISAM and you should see a decent increase in insertion speed. You’ll have to give up your foreign key constraint, but I’ve always regarded that as a nice-to-have.

I’ve never tried this myself, but I have a hunch that if you sort your data by primary key before loading it, you might see some benefit, though with indexes disabled, it may not matter.

Depending on what you want to do with your data after you’ve got it loaded, you may want to look into partitioning to keep the table and index sizes more manageable. It might not be necessary, because I’m sure 1.4G of text data shrinks down considerably when its stored in the db.

For me, I found that upping innodb_autoextend_increment helped my import processes considerably.

Since I was importing data SO fast, I found the db was tripping over itself making 8MB increments so very often. You could, of course, up your initial file size to say 10% over your estimated db size, would would mean that the DB doesn’t have to be constantly trying to grow your tablespace.

The huge performance drop was actually caused by wrong primary key creation. Instead of PRIMARY KEY USING BTREE (a,b,c) I wrote PRIMARY KEY USING BTREE (b,c,a)… Thanks for the hints though. I might use them to tweak up the speed even more.