9

I'm experiencing some problems with my MYSQL server installed with Cpanel. My server is showing 100% usage of mysql even if I don't have too much traffic on web.

I am running CentOS 6.5x64 with 80GB SSD & 8GB Ram with 4CPU core.

my.cnf:

[mysqld]
max_connections=500
key_buffer=128M
myisam_sort_buffer_size=35M
join_buffer_size=1M
read_buffer_size=1M
sort_buffer_size=2M
table_cache=4000
thread_cache_size=286
interactive_timeout=25
wait_timeout=7000
connect_timeout=10
max_allowed_packet=268435456
max_connect_errors=10
query_cache_limit=2M
query_cache_size=256M
query_cache_type=1
tmp_table_size=16M
innodb_buffer_pool_size=134217728
open_files_limit=10000
[mysqld_safe]
[mysqldump]
quick
max_allowed_packet=16M
[myisamchk]
key_buffer=64M
sort_buffer=64M
read_buffer=16M
write_buffer=16M

Here is the output of MySQLTuner:

-------- Storage Engine Statistics -------------------------------------------
[--] Status: +ARCHIVE +BLACKHOLE +CSV -FEDERATED +InnoDB +MRG_MYISAM
[--] Data in MyISAM tables: 164M (Tables: 88)
[--] Data in InnoDB tables: 33M (Tables: 239)
[--] Data in PERFORMANCE_SCHEMA tables: 0B (Tables: 17)
[!!] Total fragmented tables: 129

-------- Security Recommendations  -------------------------------------------
[OK] All database users have passwords assigned

-------- Performance Metrics -------------------------------------------------
[--] Up for: 15m 2s (83K q [92.549 qps], 13K conn, TX: 44M, RX: 5M)
[--] Reads / Writes: 61% / 39%
[--] Total buffers: 544.0M global + 4.5M per thread (500 max threads)
[OK] Maximum possible memory usage: 2.7G (34% of installed RAM)
[OK] Slow queries: 0% (0/83K)
[OK] Highest usage of available connections: 23% (116/500)
[OK] Key buffer size / total MyISAM indexes: 128.0M/59.2M
[OK] Key buffer hit rate: 97.9% (24K cached / 527 reads)
[OK] Query cache efficiency: 54.3% (18K cached / 33K selects)
[OK] Query cache prunes per day: 0
[OK] Sorts requiring temporary tables: 1% (5 temp sorts / 388 sorts)
[OK] Temporary tables created on disk: 0% (0 on disk / 374 total)
[OK] Thread cache hit rate: 99% (116 created / 13K connections)
[OK] Table cache hit rate: 98% (465 open / 474 opened)
[OK] Open file limit used: 3% (311/10K)
[!!] Table locks acquired immediately: 58%
[OK] InnoDB buffer pool / data size: 128.0M/33.6M
[OK] InnoDB log waits: 0
-------- Recommendations -----------------------------------------------------
General recommendations:
Run OPTIMIZE TABLE to defragment tables for better performance
MySQL started within last 24 hours - recommendations may be inaccurate
Enable the slow query log to troubleshoot bad queries
Optimize queries and/or use InnoDB to reduce lock wait

Here is the output of my TOP:

top - 12:01:06 up 1 day,  1:02,  1 user,  load average: 10.37, 9.22, 8.61
Tasks: 399 total,   1 running, 397 sleeping,   0 stopped,   1 zombie
Cpu(s): 84.1%us, 15.1%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.2%si,      0.7%st
Mem:   8191528k total,  6042780k used,  2148748k free,   242064k buffers
Swap:        0k total,        0k used,        0k free,  4785012k cached

PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
24515 mysql     20   0 2732m 165m 6264 S 303.8  2.1  42:43.18 mysqld
19471 root  20   0  174m  10m 6148 S  1.0  0.1   0:00.03 php
19523 root  20   0  172m  10m 5864 S  1.0  0.1   0:00.03 php
19533 root  20   0  172m   9m 5804 S  0.7  0.1   0:00.02 php
19535 root  20   0  172m  10m 5864 S  0.7  0.1   0:00.02 php

What could be the cause of this problem?
Do I have to make changes to my my.cnf to prevent server from hanging?
Or do I need to buy a new server with 16GB ram?

2
  • What is the output of the sql command show full processlist;? (Add it to the question)
    – wurtel
    Apr 30 '15 at 7:52
  • Off course, changes in my.cnf as well as RAM upgrade will definitely help
    – pikai
    Apr 30 '15 at 19:59
18

Looking at your details, we can conclude these general recommendations :

  1. Your READ : WRITE ratio (61 : 39 ) gives the clue that your engine types should be changed to Innodb. Before that check the output of

    SHOW ENGINE INNODB STATUS\G; 
    SHOW FULL PROCESSLIST;
    

    during peak hours. Most probably you will see queries in LOCK state. Table locks is the one big limitation of MyISAM. If you have a huge number of concurrent writes and selects and query performance must be consistently fast, Innodb is the only choice due to a better locking mechanism (row-level)

  2. Total fragmented tables: 129. Run OPTIMIZE TABLE and defragment all for better performance

  3. Have a look in your slow_query.log , Check Query_time, Lock_time, Rows_sent, Rows_examined. If you find queries where ratio of Rows Sent / Rows Examined set is high, then those query is good for optimization. It may be possible that some queries in slow_query.log is utilizing a lot of CPU resources.

  4. Check the status of server regarding RAM and available disk space: free -m , df -h. If you have memory available, increase the innodb_buffer_pool_size so that there is more caching.

  5. Always good to tune these parameters in MySQL configuration

    innodb_fast_shutdown=0 innodb_log_buffer_size innodb_log_file_size innodb_flush_method=O_DIRECT query_cache_size=0


CAVEAT:

  • Checklist 1: CPU of MySQL Processes, if possible to increase CPU

  • Checklist 2: RAM, If data set is fitting into the RAM.

  • Checklist 3: Connection related parameters ABORTED_CONNECTS, CONNECTIONS, MAX_USED_CONNECTIONS, OPEN_FILES, OPEN_TABLES, OPENED_FILES, OPENED_TABLES, QUERIES, SLOW_QUERIES, THREADS_CONNECTED, THREADS_CREATED, THREADS_RUNNING, PERFORMANCE_SCHEMA, QUERY_CACHE_SIZE, WAIT_TIMEOUT, LONG_QUERY_TIME, INTERACTIVE_TIMEOUT, CONNECT_TIMEOUT, MAX_CONNECTIONS, LOCK_WAIT_TIMEOUT, TX_ISOLATION, MAX_ALLOWED_PACKET, GENERAL_LOG, TABLE_OPEN_CACHE, SLOW_QUERY_LOG, OPEN_FILES_LIMIT, INNODB_LOCK_WAIT_TIMEOUT

  • Checklist 4: Memory related parameters key_buffer_size, query_cache_size, innodb_buffer_pool_size,innodb_log_buffer_size, max_connections, read_buffer_size, read_rnd_buffer_size, sort_buffer_size, join_buffer_size, binlog_cache_size, thread_stack, tmp_table_size

  • Checklist 5: Other Important InnoDB parameters innodb_log_file_size ,innodb_file_per_table ,innodb_flush_log_at_trx_commit ,innodb_flush_method ,innodb_fast_shutdown

  • Checklist 6: mysqltuner

    Joins performed without indexes: Table cache hit rate:

    -------- MyISAM Metrics -------------------------------------------- 
    [!!] 
    [!!] Write Key buffer hit rate: 
    -------- InnoDB Metrics -------------------------------------------- 
    [!!] InnoDB buffer pool / data size: 
    [!!] Ratio InnoDB log file size / InnoDB Buffer pool size : 
    [!!] InnoDB buffer pool instances:
    
  • Checklist 7: Processlist The queries below are found in the state of ‘Sending data’ for long time

  • Checklist 8: Slow query log - Rows_examined vs Rows_sent ratio very high,

  • Checklist 9: duplicate indexes (if any)

  • Checklist 10: explain plan of few slow queries

  • Checklist 11: Check if where clause columns are properly indexed (non-indexed queries)

  • Checklist 12: Check for table lock, metadata lock, deadlock if any.

  • Checklist 13: Any maintenance job is running in intervals (eg, mysqlcheck, mysqldump etc)

  • Checklist 14: Hardware resources, slow disks, RAID rebuilding, disk I/O, saturated network, network bandwidth w.r.t throughput with increase in number of threads.

1
  • 4
    Could you explain the reasoning behind making each of the changes in you suggested in item number 5 above? Or, at the very least, perhaps the benefit of setting each value? Jul 1 '16 at 15:06

Not the answer you're looking for? Browse other questions tagged or ask your own question.