Re: Question about Maglev algorithm

2018-12-29 Thread Valentin Vidic
On Sat, Dec 29, 2018 at 06:03:51PM +0100, Aleksandar Lazic wrote:
> I thought I have misunderstood the Idea behind maglev, thanks for 
> clarification.

Found another mention of Maglev [Eis16] for high-level load balancing (between
datacenters):

  https://landing.google.com/sre/sre-book/chapters/load-balancing-frontend/

-- 
Valentin



Re: Question about Maglev algorithm

2018-12-28 Thread Valentin Vidic
On Fri, Dec 28, 2018 at 07:11:24PM +0100, Aleksandar Lazic wrote:
> Have anyone take a look into the Maglev algorithm ?
> 
> This paper looks very interesting 
> https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/44824.pdf

Seems to be in IPVS already:

https://github.com/torvalds/linux/blob/master/net/netfilter/ipvs/ip_vs_mh.c

-- 
Valentin



Re: [PATCH] CLEANUP: Fix typos in the proto_tcp subsystem

2018-11-25 Thread Valentin Vidic
On Sun, Nov 25, 2018 at 11:30:06AM -0800, Joseph Herlant wrote:
> -/* Returns some tcp_info data is its avalaible. "dir" must be set to 0 if
> - * the client connection is require, otherwise it is set to 1. "val" 
> represents
> +/* Returns some tcp_info data is its available. "dir" must be set to 0 if
> + * the client connection is required, otherwise it is set to 1. "val" 
> represents

Should this sentence read:

  Returns some tcp_info data if it's available.

-- 
Valentin



Re: Why is there a tilde ~ character behind the frontend name in the log file?

2018-02-07 Thread Valentin Vidic
On Wed, Feb 07, 2018 at 10:15:56AM +, Pieter Vogelaar wrote:
> I have a http frontend “default-http” and “default-https”. In the
> access log is the ~ (tilde) character appended to the default-https
> frontend name, like “default-https~”.

Looks like it means SSL:

  |   | %ft  | frontend_name_transport ('~' suffix for SSL)  | string |

-- 
Valentin



Haproxy sessions vs connections

2016-09-05 Thread Valentin Vidic
Hi,

I have a haproxy in front of a couple apache servers in prefork
mode.  Looking at the haproxy stats page shows only a couple of
active sessions per server while the apache status shows a
larger number of requests in keep-alive state.  Is there an
option that would make haproxy count keep-alive connections
as active sessions too?

Without this maxconn limit I set on servers does not seem to
make sense.  For example:

maxconn per server: 1000
keep-alive connections: ~100
active sessions: ~10

-- 
Valentin