Fails redis benchmark
Toxiproxy responds with conn reset when running redis benchmark, and redis-cli hangs intermittently.
Steps to reproduce: create a proxy to forward to a real redis run redisbenchmark
redis-benchmark -p 53002
...
99.97% <= 46 milliseconds
99.98% <= 48 milliseconds
99.99% <= 50 milliseconds
100.00% <= 52 milliseconds
22573.36 requests per second
Error: Connection reset by peer: 78.47770
server version is 2.1.0
Hey @qzaidi I failed to reproduce this on my machine. (Mid 2015 Mac Book Pro)
./toxiproxy-server # start toxiproxy
./redis-server # start redis
./toxiproxy-cli new redis -l localhost:7777 -u localhost:6379 # add a redis proxy
./redis-benchmark -p 7777 # run the benchmark
redis-benchmark ran to completion without error.
Any chance you could provide more information about your failure? How do you know Toxiproxy is at fault rather than redis?
Hi @jpittis
I ran the above on an AWS instance (ubuntu 14.04). I think toxiproxy is at fault because a similar benchmark against an actual redis instance (no toxiproxy in between) worked. Actually I suspected toxiproxy after our application clients would intermittently get conn resets without any toxics, and tried to see if i can make it easier to reproduce, hence the benchmark.
I will try to get more details about this (and will try it on my mac as well).