Le 10/04/2019 à 08:44, Kevin Zhu a écrit :


---------- Forwarded message ---------
From: *Kevin Zhu* <[email protected] <mailto:[email protected]>>
Date: Wed, 10 Apr 2019 at 14:25
Subject: Re: haproxy1.9, SPOA: too many open files
To: Christopher Faulet <[email protected] <mailto:[email protected]>>


Thinks reply.

OS: CentOS Linux release 7.4
HW: platform: KVM;  CPU: Intel Xeon E3-12xx v2 (Ivy Bridge) * 1;   mem: 2048M

HAProxy Make: make TARGET=linxu2628 USE_THREAD=1
SPOA: 1:cd contrib/spoa_example ;
            2: set async and pipelining and fragmentation = true in souces file spoa.c
             3: make
             4:./spoa

Benchmark: see abtest.sh
Configurations files: is attached in this mail

I'm testing results(result value l is approximately):
HAProxy-1.8.16:
         multi procs single thread:
                 CPU: HAProxy: 10% * 8; spoa: 4% - 7%
                TCP HAProxy and spoa(netstat -anpt | grep 12345 | grep EST | wc -l): 150 - 400
                 Time taken for tests:   67 seconds

         multi threads single thread:
                 CPU: HAProxy: 85% - 90%; spoa: 3% - 5%
                TCP HAProxy and spoa(netstat -anpt | grep 12345 | grep EST | wc -l): 100 - 200
                 Time taken for tests:   53 seconds

HAProxy-1.9.6:
         multi procs single thread:
                 CPU: HAProxy: 9% * 8; spoa: 24% - 27%
                TCP HAProxy and spoa(netstat -anpt | grep 12345 | grep EST | wc -l):8050
                 Time taken for tests:   99 seconds

         multi threads single thread:
                 CPU: HAProxy: 80%; spoa: 9% - 13%
                TCP HAProxy and spoa(netstat -anpt | grep 12345 | grep EST | wc -l): 7500 - 10200
                 Time taken for tests:   65 seconds


Hi Kevin,

Sorry for the delay, I was pretty busy on other bugs.

So, first of all, I'm surprised. Looking at your test file (abtest.sh), you send 10k requests with 1000 concurrent connections in ~60 secs or more. It's not really fast. And the workload remains low. So there is probably a bottleneck somewhere.

Could you enable the logs on HAProxy, both for the HTTP and the SPOE, and check if there is something strange. Maybe timeouts or errors. Or a timer with an high value. If possible, share it. I'm interested to take an eye on it.

You could also start the spoa in debug mode. It is pretty verbose, but it can help to find the problem.

Could you also use latest 1.8 and 1.9 snapshots ?

--
Christopher Faulet

Reply via email to