0

Here is my my.cnf

symbolic-links=0
user            = mysql
pid-file        = /var/run/mysqld/mysqld.pid
socket          = /var/run/mysqld/mysqld.sock
basedir         = /usr
datadir         = /var/lib/mysql
tmpdir          = /tmp
log-error = /var/log/mysql/error.log
log-warnings = 2
symbolic-links=0
lc-messages-dir = /usr/share/mysql
skip-external-locking
key_buffer_size = 2G
thread_stack            = 192K
thread_cache_size       = 350
myisam-recover         = BACKUP
max_connections        = 1500
query_cache_limit       = 100M
query_cache_size        = 512M
general_log             = 1

long_query_time = 10
bind-address = 0.0.0.0
old_passwords=1
skip-secure-auth

innodb_buffer_pool_size  = 30G
innodb_stats_on_metadata=0
table_open_cache = 4096
skip-external-locking
skip-name-resolve
ft_min_word_len = 2
innodb_buffer_pool_instances = 25 
innodb_file_per_table=0
innodb_checksum_algorithm=INNODB
binlog_checksum=NONE

24 Core + 60GB of RAM + RAID controller + 4 TB dataset in MySQL-- mostly MySQL with some INNODB

and MySQL server starts consuming swap space, even when there is about 40GB RAM available. This is a production system with an average of 1000 connections during business hours

This is a Debian-9.12 Stretch Dell-R320 System running only MySQL Database server.

top/htop both show the server is in pain because swap ( 4GB ) has been fully used and no further operations are smoother.

I have to restart the server every week to make it clear from swap and so on.

I wanted to increase the RAM usage but its relying on SWAP

vm.swappiness=0

David Lee
  • 665
  • 7
  • 20
Nasir Mahmood
  • 161
  • 3
  • 7

1 Answers1

0

You could also find a balance between vm.swapiness and vm.vfs_cache_pressure.

Maybe OS is caching a lot, and it could be the reason of swap memory reclaim.

Try to low vm.vfs_cache_pressure if helps.

Pablo Luna
  • 369
  • 3
  • 5