Hi Everyone,
I’m having a problem with a single repair query seemingly taking over the whole server. I’ll try and explain…. Large table or 1.2m rows, running a repair on this causes the query queue to grow and grow and finally max out. The problem isn’t table locks because no-one else is using the test database I am repairing. The live database is the one with all the queries coming in. This is confirmed with mytop. It just seems the single repair is munching up all the system ram and not allowing any other queries a share. Even the most simple select queries on the live database are taking 500 secs!! I want the repair to run but not to the detriment of other queries. Once the repair is over it can take about 3 minutes before the queue starts to evaporate also, shouldn’t it kick back into gear as soon as the repair is over? We are on myisam tables, MySQL 5.0 on an 8Gb 32bit Debian system (large memory kernel installed) I’ve tried playing with myisam_* settings in my.cnf but it doesn’t seem to work, Here is a snip of my current my.cnf ----------------- skip-locking skip-innodb skip-bdb key_buffer = 2G max_allowed_packet = 16M table_cache = 1500 thread_cache_size = 200 thread_concurrency = 8 sort_buffer_size = 8M read_buffer_size = 2M join_buffer_size= 8M read_rnd_buffer_size = 16M tmp_table_size=64M #INDEX CREATION myisam_sort_buffer_size = 64M #default 8M myisam_max_sort_file_size = 1G #default 2G myisam_max_extra_sort_file_size = 2G myisam_repair_threads = 1 #QUERY CACHE query_cache_limit = 1M query_cache_size = 128M query_cache_min_res_unit = 1024 query_cache_type = 1 #SLOW LOG log-slow-queries = /var/log/mysql/mysql-slow.log log_long_format long_query_time=5 #FULL TEXT SETTINGS ft_min_word_len = 2 ft_max_word_len = 20 #CONNECTIONS max_connections = 250 max_connect_errors = 100 open-files-limit=4096 ----------------- Thanks in advance. Gary. -- No virus found in this outgoing message. Checked by AVG Free Edition. Version: 7.5.446 / Virus Database: 268.18.7/713 - Release Date: 07/03/2007 09:24
