Hi, mysqltuner does only take some buffers to calculate the maximum buffer size, for example it does not take in account the Aria table cache or the TokuDB table cache (if you have TokuDB enabled) or other MariaDB specific buffers it more to give you an idea of the possible memory usage than a real value. ps: you could be having a problem with your Maria but without more informations its hard to know if something is not right. Regards. On 10/02/2014 15:57, ImperialNetwork wrote:

Hi,

Running a mysqltuner.pl scan tells the following: [--] Up for: 6d 15h 31m 3s (139M q [242.402 qps], 2M conn, TX: 427B, RX: 16B) [--] Reads / Writes: 95% / 5% [--] Total buffers: 2.0G global + 2.3M per thread (175 max threads) [OK] Maximum possible memory usage: 2.4G (31% of installed RAM)

However , it seems to be using more than what mysqltuner reports (3.3G now):

[~]# ps -eo comm,rss|awk '{arr[$1]+=$2} END {for (i in arr) {print arr[i]/1024, i}}'|grep -v '^0 '|grep mysqld 3314.3 mysqld

my.cnf:

[mysqld] user = mysql extra_port = 3308 default_storage_engine = MyISAM sql-mode = NO_ENGINE_SUBSTITUTION

# DATA STORAGE # datadir = /databases/mysql

# MariaDB # thread_handling = pool-of-threads optimizer_switch ='mrr=on' optimizer_switch ='mrr_sort_keys=on' optimizer_switch ='mrr_cost_based=off' optimizer_switch ='join_cache_incremental=on' optimizer_switch ='join_cache_hashed=on' optimizer_switch ='join_cache_bka=on' mrr_buffer_size = 1M aria_pagecache_buffer_size = 128M aria_sort_buffer_size = 128M key_cache_segments = 8

# MyISAM # key_buffer_size = 256M myisam_recover = BACKUP,FORCE myisam_sort_buffer_size = 128M group_concat_max_len = 4096 max_seeks_for_key = 1000 #myisam_use_mmap = 1

# SAFETY # max_allowed_packet = 32M max_connect_errors = 1000 local-infile = 0

# CACHES AND LIMITS # tmp_table_size = 128M max_heap_table_size = 128M query_cache_type = 1 query_cache_size = 128M query_cache_limit = 1M query_cache_min_res_unit = 2K max_connections = 175 max_user_connections = 75 thread_cache_size = 32 open_files_limit = 50K table_definition_cache = 10K table_open_cache = 40K join_buffer_size = 1M sort_buffer_size = 512K read_buffer_size = 256K read_rnd_buffer_size = 256K

# INNODB # sysdate_is_now = 1 innodb = FORCE innodb_strict_mode = 1 innodb_flush_method = O_DIRECT innodb_log_files_in_group = 2 innodb_log_file_size = 256M innodb_flush_log_at_trx_commit = 2 innodb_file_per_table = 1 innodb_buffer_pool_size = 1536M innodb_read_io_threads = 4 innodb_write_io_threads = 4 innodb_io_capacity = 400

# LOGGING # slow_query_log = 1 long_query_time = 5

connect_timeout=30 interactive_timeout=120 wait_timeout=120

Is this some memory leak problem , problems with the script or normal behavior ?

Thank you in advance, Alexandru