Why redis-py causes high cpu?

Asked Nov 16 '15 at 00:52

Active Nov 16 '15 at 08:37

Viewed 528 times

My website is powered by tornado. When doing the stress testing, I found python3 used almost all cpu while redis-server used a little. So, I did a profile, but the profile result makes me confused:

It seems that all redis-py operations (r.xxx() and lua_xxx()) takes huge cpu time.

I have no idea why python3 eats all cpu since if the redis request is blocked, it should sleep thus less cpu usage. So, could you please tell me where could the problem be? Thanks!

edited Nov 16 '15 at 08:37

Nick Bondarenko

6,211
4
35
56

asked Nov 16 '15 at 00:52

abcdabcd987

2,043
2
23
34

The CPU time seems to be used by `lua_place_order()` but why is hard to say without it's code. – Klaus D. Nov 16 '15 at 01:44
@KlausD. `lua_place_order()` is a previously registered redis lua script, so I think it should be executed at redis-server, thus does not make python crazy. – abcdabcd987 Nov 16 '15 at 01:48
you can get a 17% increase in speed of the string manipulation using .format() http://stackoverflow.com/a/20746724/1499803 – Back2Basics Nov 16 '15 at 01:53
@Back2Basics Thanks for the tip! – abcdabcd987 Nov 16 '15 at 02:02

Why redis-py causes high cpu?

0 Answers0