go-cockroach

README

go-cockroach

Scrappy load generation tool for cockroach via HTTP.

• Heavy run load 11K INSERTS 100K SELECT Target per transaction (Stability / Schema run).
• go get github.com/lib/pq
• go get github.com/gorilla/mux
• go get github.com/Pallinder/go-randomdata
• go get github.com/geekzoo/go-cockroach
• cd $GOPATH/github.com/geekzoo/go-cockroach/
• vim cal.go
  Change host to your cockroachdb host name and port, user name/password
  I use HAProxy for load balancing to the back end database server, I'll include a skel config at some point.
  
• go build && ./go-cockroach
• curl -s http://domain.com:8000/cal_prep #creates db and table
• ab -c 100 -n 1000 -k http://domain.com:8000/cal_insert
• ab -c 100 -n 1000 -k http://domain.com:8000/cal_all
• in a web browser goto
• http://domain.com:8000/cal_insert
• http://domain.com:8000/cal_all
• http://domain.com:8000/show_sessions

TODO:

• Add runtime options for http listen port
• Add connection to carbon and influx
• Add Base HAProxy config
• Add condition for nil when a cockroach node is down /show_sessions!!!

BUGS:

• show_sessions Panic if cockroach node is down and return nil
• Tcc or stdout carbon returns metric's per sql call N time per second. default aggration is every 10 seconds resulting in loss of resolution unless changed in carbon even if changed to every second we still lose N per seconds.

Documentation

Overview

* Heavy run load 11K INSERTS 100K SELECTS per second (Stability / Schema run). * curl -s http://ipORdomain:8000/cal_prep #creates db and table * Scraper v_2 go-bench-cal_insert * ab -c 100 -n 1000 -k http://domain.com/cal_insert * ab -c 100 -n 1000 -k http://domain.com/cal_all * in a web browser goto * http://domain.com/cal_prep * http://domain.com/cal_insert * http://domain.com/cal_all