Mailing-List: contact users-help@tomcat.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tomcat Users List" <users@tomcat.apache.org>
Received-SPF: pass (nike.apache.org: domain of knobisoft@gmail.com designates
 209.85.223.173 as permitted sender)
MIME-Version: 1.0
In-Reply-To: 
 <CAJtcoLauF2BvedGKHRfuPYUfB8NymmFaL3E_nAUOSdhrEfw7vQ@mail.gmail.com>
References: 
 <CAJtcoLauF2BvedGKHRfuPYUfB8NymmFaL3E_nAUOSdhrEfw7vQ@mail.gmail.com>
Date: Tue, 1 Jul 2014 18:01:54 +0200
Message-ID: 
 <CAJtcoLaZWDCW0PKNzd=OYsUmY4up1-pw4eY_wNz0nabpm1YgMw@mail.gmail.com>
Subject: Re: jkmanager: how to interpret "Current number of backend
 connections"
From: Martin Knoblauch <knobisoft@gmail.com>
To: Tomcat Users List <users@tomcat.apache.org>
Content-Type: multipart/alternative; boundary=089e0111be1601b31e04fd23e5d7

--089e0111be1601b31e04fd23e5d7
Content-Type: text/plain; charset=UTF-8

On Mon, Jun 30, 2014 at 5:18 PM, Martin Knoblauch <knobisoft@gmail.com>
wrote:

> Hi,
>
>  our environment is (I know, some are old ...):
>
> Apache 2.2.21
> mod_jk-1.2.40
> Tomcat 7.0.42
>
> My question is how to interpret the numbers from "jkmanager", expecially
> the Busy/Max/Con numbers.
>
>
> "Busy" and "Max" make sense to me. "Busy" is always less than "Max" and
> "Max" gets bumped up from time to time. Matches the description.
>
> Now, "Con" seems to grow steadily over time and is not reset when we do a
> manual "Reset" of the worker status. Somehow this does not match the
> description of "Current number of backend connections". At least not for me
> :-)
>
> Cheers
> Martin
>
>
OK, seems the continuous growing  of the "Current number of backend
connections" numbers is related to having timeouts configured. It also
seems to have started between mod_jk-1.2.32 and mod_jk-1.2.37. I can see
some heavy redo of the handling of the "connected" counter.

In any case, the following configuration seems to trigger the behaviour:

worker.template.type=ajp13
worker.template.lbfactor=1
worker.template.prepost_timeout=15000
worker.template.connect_timeout=15000
worker.template.ping_timeout=15000
worker.template.ping_mode=A
worker.template.connection_ping_interval=180
worker.template.retries=3
worker.template.connection_pool_timeout=30

Without setting the timeout values, the displayed number of backend
connections seems to be stable. I suspect a missing decrement somewhere,
just hard to say where.

Cheers
Martin
-- 
------------------------------------------------------
Martin Knoblauch
email: k n o b i AT knobisoft DOT de
www: http://www.knobisoft.de

--089e0111be1601b31e04fd23e5d7--