First time i am sorry about my English , but please help me on my.cnf . I read many post about config Mysql for best optimize but not thing to imporve. My server is Data center that contains two windows and linux . window has 5 core 5 GBram and linux has 18 CPU , 20GB ram . My window has two software that was developed by c# and i am using dll mysql.data . two software ,they will craw data from another website and insert into data to linux , sometime call "select , update , delete" . they work every time about 10s to craw and 10s to craw and send data to linux . On my linux with 358 user online same time , they click on link . another link and stay very long , they call always call to Database every time . My server has 60% select and 40% write . I dont know how to config my.cnf . But some time software in window insert data to table of database on linux very long about 30s and some time it show not respoding . I dont know why and i dont know what parameter make table innodb lock or make innodb can not be inserted by c# . Now i would like to know how can i change and what parameter(variable) to change to get best performance for select and insert into with larger data . Thank you very much . This is my my.cnf ( i am using mysql 5.6.31 – MySQL Community Server (GPL))
[mysqld] datadir=/var/lib/mysql socket=/var/lib/mysql/mysql.sock user=mysql skip_name_resolve skip_external_locking default_storage_engine=InnoDB max_connections=20000 join_buffer_size=2M open_files_limit=65535 query_cache_type=ON query_cache_size=1G thread_concurrency=36 thread_cache_size=25 table_open_cache=10048 wait_timeout=300 read_buffer_size=1M sort_buffer_size=1M read_rnd_buffer_size=1M max_connect_errors=1000 long_query_time=7 key_buffer_size=128M max_allowed_packet=268435456 interactive_timeout=400 myisam_sort_buffer_size = 64M tmp_table_size=64M max_heap_table_size=64M innodb_file_per_table innodb_flush_method=O_DIRECT innodb_log_file_size=512M innodb_buffer_pool_size=10G innodb_buffer_pool_instances=10 innodb_log_buffer_size=64M innodb_thread_concurrency=0
Best Answer
On 100
It is not practical, nor efficient, to try perform 100 queries simultaneously. Because of that, there is no way to "configure" it into existence. It is better to do...
INSERTsare into the same table, then use a single "batch"INSERTwith 100 rows. This will run 10 times as fast as a 100 1-rowINSERTs.If it is not 100 similar inserts, please provide more details about the queries.