Please report any other irregularities here.
The SCL3 instances are actually using a custom MySQL config, that differs slightly from the MySQL 5.6 defaults. Most of the settings involve paths/buffer sizes/other settings that are not relevant on RDS, however there are some that are.
Created attachment 8796899 [details] SCL3 MySQL config MySQL settings docs: http://dev.mysql.com/doc/refman/5.6/en/server-options.html http://dev.mysql.com/doc/refman/5.6/en/innodb-parameters.html http://dev.mysql.com/doc/refman/5.6/en/fulltext-fine-tuning.html https://dev.mysql.com/doc/refman/5.6/en/replication-options-slave.html AWS parameter group settings (default profile is 'default.mysql5.6'): https://console.aws.amazon.com/rds/home?region=us-east-1#parameter-groups:id=
Relevant settings we should probably keep: character_set_server=utf8 -> default is 'latin1' (http://dev.mysql.com/doc/refman/5.6/en/server-options.html#option_mysqld_character-set-server) collation_server=utf8_bin -> default is 'latin1_swedish_ci' (http://dev.mysql.com/doc/refman/5.6/en/server-options.html#option_mysqld_collation-server) slow_query_log = 1 -> default is 0 ie off (http://dev.mysql.com/doc/refman/5.6/en/server-system-variables.html#sysvar_slow_query_log) long_query_time=2 -> default is 10 seconds (http://dev.mysql.com/doc/refman/5.6/en/server-system-variables.html#sysvar_long_query_time) We don't need to keep 'ft_min_word_len=2' since that only applies to MyIASM tables: http://dev.mysql.com/doc/refman/5.6/en/server-system-variables.html#sysvar_ft_min_word_len http://dev.mysql.com/doc/refman/5.6/en/fulltext-fine-tuning.html
Created attachment 8796900 [details] [review] devservices-aws PR #2
Looks good. Merged and applied. aws_db_instance.treeherder-heroku: Modifying... parameter_group_name: "default.mysql5.6" => "treeherder" aws_db_instance.treeherder-stage-rds: Modifying... parameter_group_name: "default.mysql5.6" => "treeherder"
Many thanks! I've rebooted the two RDS instances to pick up the slow query log changes, since they aren't dynamic settings. I used the failover option on one to compare: with failover reboot took 55 seconds, without 11 seconds. Not bad either way :-)
Status: ASSIGNED → RESOLVED
Last Resolved: 2 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.