io - Heavy replication write load on MySQL slave -


we using percona mysql 5.6 on debian 8 ecommerce aggregator. there master backend server doing whole etl (processing product feeds partners) , slave mysql server used frontend web servers. it's single product database 600gb data.

both machines have raid10 datacenter series ssds. master mysql dual xeon e5 128gb ram , slave single xeon e5 64gb ram. our problem is, etl i/o heavy (with thousands of iops), master able handle i/o load slave server cannot keep replication. work done in ramdisk , real neccessary data written database, already.

the slave has more reads writes, replication cannot catch up. options there scaling replication writes (i.e. i/o load) on slave?

edit 2016-11-18: "options" i'm not asking optimizing mysql, other techniques or software handle situation better.

upgrading 5.7 should solve immediate problem: http://mysqlhighavailability.com/multi-threaded-replication-performance-in-mysql-5-7/.

this precursor master running out of write capacity. should start thinking sharding soon. can check out youtube's vitess.io project, i'm part of.


Comments

Popular posts from this blog

Formatting string according to pattern without regex in php -

c - zlib and gdi32 with OpenSSL? -

java - inputmismatch exception -