How to fix high memory usage of MariaDB?
I am using MariaDB(10.1.21) with following storage engines:
- InnoDB
- TokuDB
- Spider Engine
System Configuration is :
- 3GB RAM
- Dual Core Processor
I have tried pt-mysql-summary tool to identify the memory usage and it shows 90% of innodb buffer pool size is filled,when no process is running on server.
I have also tried restarting the MariaDB server and this reduces memory for about 1-2 hours only and after that it eats up memory.
Any clue why it is consuming too much memory?
Sharing my configurations below:
My.cnf
[client]
port = 3306
socket = /var/run/mysqld/mysqld.sock
# Here is entries for some specific programs
# The following values assume you have at least 32M ram
# This was formally known as [safe_mysqld]. Both versions are currently parsed.
[mysqld_safe]
socket = /var/run/mysqld/mysqld.sock
nice = 0
[mysqld]
#performance_schema=ON
tmp_table_size=16M
max_heap_table_size=16M
skip-log-bin
#table_open_cache =16384
#table_definition_cache =16384
#
# * Basic Settings
#
user = mysql
pid-file = /var/run/mysqld/mysqld.pid
socket = /var/run/mysqld/mysqld.sock
port = 3306
basedir = /usr
datadir = /var/lib/mysql
tmpdir = /tmp
lc_messages_dir = /usr/share/mysql
lc_messages = en_US
skip-external-locking
#bind-address = 127.0.0.1
#skip-networking
# SAFETY #
max_allowed_packet= 16M
max-connect-errors = 1000000
skip-name-resolve
#sql-mode = STRICT_TRANS_TABLES,ERROR_FOR_DIVISION_BY_ZERO,NO_AUTO_CREATE_USER,NO_AUTO_VALUE_ON_ZERO,NO_ENGINE_SUBSTITUTION,NO_ZERO_DATE,NO_ZERO_IN_DATE,ONLY_FULL_GROUP_BY
# CACHES AND LIMITS #
tmp-table-size = 32M
max-heap-table-size = 32M
query-cache-type = 0
query-cache-size = 0
max-connections = 500
thread-cache-size = 50
open-files-limit = 65535
table-definition-cache = 1024
table-open-cache = 2048
# LOGGING #
log_warnings = 2
slow_query_log =1
log-output = TABLE
long_query_time = 5
log_slow_verbosity = query_plan
log-error =/var/log/mysql/system_error.err
# INNODB #
innodb-flush-method = O_DIRECT
innodb-log-files-in-group = 2
innodb-log-file-size = 128M
innodb-flush-log-at-trx-commit = 1
innodb-file-per-table = 1
innodb-buffer-pool-size = 1720M
[mysqldump]
quick
quote-names
max_allowed_packet = 16M
[mysql]
#no-auto-rehash # faster start of mysql but no tab completion
[isamchk]
key_buffer = 16M
TokuDB.cnf
plugin-load-add=ha_tokudb.so
tokudb_data_dir=/var/lib/mysql/toku_db_data
tokudb_read_block_size=64k
tokudb_row_format=tokudb_zlib
tokudb_directio=on
Spider.cnf
[mariadb]
#spider_internal_limit =1
spider_direct_order_limit =1
spider_skip_default_condition =1
spider_casual_read =1
spider_bgs_mode =2
spider_direct_dup_insert =1
spider_auto_increment_mode =2
#optimizer_switch='engine_condition_pushdown=on'
#optimizer_switch='mrr=on,mrr_sort_keys=off'
Answer
innodb-buffer-pool-size = 1720M is much too large for a 3GB machine, especially if TokuDB and Spider are also in use.
The real question is, "is there any swapping"? If there is no swapping, then there is no problem. If there is swapping, then the first defense is to lower that setting. I suggest 500M. Swapping is very bad for performance of MySQL.
It is normal to see 90%. 92% is the median value for (Innodb_buffer_pool_bytes_data / innodb_buffer_pool_size) a sampling of machines. Since the buffer_pool is a cache, it is normal to be "full" most of the time. (It rarely gets over 98% full -- probably leaving some space for quick reuse.) If it is below, say, 50% full after running for several hours, I would say the innodb_buffer_pool_size is too big for the dataset in question.
Suggest lowering to long_query_time = 2 in order to catch faster, but frequent, queries, too.