I’m using a replication slave for backup purposes, application is not slave aware and can’t be changed easily. The master runs about 60 queries/s, slave is about 3 queries/s.
I will usually copy master’s my.cnf and cut down a few vars, max_connections/mas_user_connections to 10, key_buffer to 32MB, sort/read/query/innodb buffers 2-3x smaller than master. I keep exactly the same innodb log file sizes )
It will usually keep at 0s behind master but this particular server is always behind, it just can’t catch up.
It has 2 databases live and staging, 140 InnoDB tables, 2-3 MyISAM (small ones), used for sequence emulation, interprocess communication inside transactions, logs.
I’m trying to change vars but with no results so far. Any hints at what should I try next? Master has faster disks.
Also in general what config should I use for a slave that’s used only for backup?