In 2.5.3 with a fresh MariaDB install, innodb_log_buffer_size increase message emits

[davidfavor]

MariaDB version I'm running...

lxd: net15-template-noble # mariadb --version
mariadb from 11.3.2-MariaDB, client 15.2 for debian-linux-gnu (x86_64) using  EditLine wrapper

Ran innodb_log_buffer_size up to 8G + message still emits.

Might be something to fix in a future version.

[davidfavor]

/etc/my.cnf contents...

[client]

port        = 3306
socket      = /var/run/mysqld/mysqld.sock

[mysqld_safe]

socket      = /var/run/mysqld/mysqld.sock
nice        = 0

[mysqld]

# Management + Access

user               = mysql
pid-file           = /var/run/mysqld/mysqld.pid
basedir            = /usr
datadir            = /var/lib/mysql
tmpdir             = /tmp
socket             = /var/run/mysqld/mysqld.sock
port               = 3306
lc_messages_dir    = /usr/share/mysql
lc_messages        = en_US

bind-address       = 127.0.0.1

# This throws many warnings, which should be resolved prior to using
#skip-name-resolve
skip-external-locking
skip-networking

performance_schema = ON

#
# * Character sets
#

# MySQL/MariaDB default is Latin1, but in Debian we rather default to the full
# utf8 4-byte character set. See also client.cnf
character-set-server  = utf8mb4
collation-server      = utf8mb4_general_ci

# Logging

log_error           = /var/log/mysql/mariadb-error.log
log_warnings        = 2

general_log_file    = /var/log/mysql/mariadb-general.log
general_log         = OFF

slow_query_log_file = /var/log/mysql/mariadb-slow.log
slow_query_log      = ON
long_query_time     = 10
log_slow_rate_limit = 1
log_slow_verbosity  = query_plan,innodb,explain
log_slow_admin_statements

# Disable by default, as wp_options has no index
# log-queries-not-using-indexes

# General Settings

# thread_cache_size max + max_connections: must match, to effect either
max_connections     = 128
thread_cache_size       = 128

connect_timeout     = 5
wait_timeout        = 600
max_allowed_packet  = 16M
sort_buffer_size    = 4M
bulk_insert_buffer_size = 16M

# Lower of these two values used, so best keep them equal
tmp_table_size      = 256M
max_heap_table_size = 256M

# General

query_cache_type           = DEMAND
query_cache_strip_comments = 1

# disable query cache: query_cache_limit + query_cache_size == 0 for high speed InnoDB

query_cache_limit          = 0
query_cache_size           = 0

# join_buffer_size: allocated for every join, so best left low, else trouble will ensue
join_buffer_size           = 10M

# MyISAM

myisam_recover_options  = BACKUP
key_buffer_size         = 128M
myisam_sort_buffer_size = 512M
concurrent_insert       = 2
read_buffer_size        = 2M
read_rnd_buffer_size    = 1M

# InnoDB

innodb_buffer_pool_size        = 256M
innodb_buffer_pool_instances   = 1

innodb_log_buffer_size         = 64M
innodb_log_files_in_group      = 2
innodb_log_file_size           = 64M

innodb_file_per_table          = 1
innodb_io_capacity             = 1000

# innodb_open_files: should be greater than table_open_cache
innodb_open_files              = 4096

# table_open_cache == 64, per http://bit.ly/1mi7c4C
table_open_cache               = 64

innodb_flush_method            = O_DIRECT
innodb_flush_log_at_trx_commit = 2

[mysqldump]
quick
quote-names
max_allowed_packet  = 16M

[mysql]
#no-auto-rehash # faster start of mysql but no tab completion

[isamchk]
key_buffer      = 16M

[PauloOF]

I'm always getting the recommendation innodb_log_buffer_size (> xM), independently of x value at mysql config.

It seems a bug.

[Neurozone]

It depends if you have innodb tables or not (which I strongly recommend) but your innodb_buffer_size is way too low