Mailing-List: contact httpclient-users-help@hc.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "HttpClient User Discussion" <httpclient-users@hc.apache.org>
Received-SPF: pass (nike.apache.org: message received from 54.76.25.247 which
 is an MX secondary for httpclient-users@hc.apache.org)
From: Tomas Tunkl <tunkl.tomas@gmail.com>
Content-Type: multipart/alternative;
 boundary="Apple-Mail=_CBB59347-1DD9-4F29-A6CF-EAE1E5901A9C"
Message-Id: <F56E8D8C-CDFA-4BCA-A884-78EA9A17578A@gmail.com>
Mime-Version: 1.0 (Mac OS X Mail 8.2 \(2098\))
Subject: Re: AbstractNIOConnPool and leasingRequests
Date: Thu, 30 Apr 2015 12:46:39 +0200
References: <E73CDD02-EB04-4703-81F8-A10465E61CE5@gmail.com>
 <1430300166.2052.0@ok2consulting.com>
 <F911818E-765A-4869-A9B8-7EDA79D8D095@gmail.com>
 <1430304115.2052.2@ok2consulting.com>
 <D22EA3C4-1D62-415C-8B97-22C1CAF9448B@gmail.com>
 <1430307965.2052.3@ok2consulting.com>
To: HttpClient User Discussion <httpclient-users@hc.apache.org>
In-Reply-To: <1430307965.2052.3@ok2consulting.com>

--Apple-Mail=_CBB59347-1DD9-4F29-A6CF-EAE1E5901A9C
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain;
	charset=us-ascii

Yes, it seems that this is my problem. I have two dumps which are 5m =
after each other. I have looked to AbstractNIOConnPool and I have in =
leasing requests 350

then I have one route only in routeToPool and there I have:

pending: 55
availble: 3
leased: 0

and these values are the same 5m after except the leasingRequests which =
are higher, but it seems that pending has not moved and the objects are =
the same. Even if I have configured SocketTimeout and Connection acquire =
timeout to 10s

When the request is removed from pending?=20

Thank you
Tom

> On 29. 4. 2015, at 13:46, Oleg Kalnichevski <olegk@apache.org> wrote:
>=20
>=20
> On Wed, Apr 29, 2015 at 1:02 PM, Tomas Tunkl <tunkl.tomas@gmail.com =
<mailto:tunkl.tomas@gmail.com>> wrote:
>> My problem is that client has stopped sending requests, and I have =
noticed that leasingRequests are not getting empty. So I was thinking =
whether this could be my problem.
>> Thank you
>> Tom
>=20
>=20
> Lease request backlog get processed every time a connection gets =
released back to the pool (there are other cases such as connect failure =
or timeout, but connection release is the primary one). If lease =
requests get piling up in the queue, that most likely means that the =
pool has been exhausted and leased connections are not being released =
back to the pool.
>=20
> Oleg
>=20
>=20
>=20
>>> On 29. 4. 2015, at 12:41, Oleg Kalnichevski <olegk@apache.org> =
wrote:
>>> On Wed, Apr 29, 2015 at 12:31 PM, Tomas Tunkl =
<tunkl.tomas@gmail.com> wrote:
>>>> Hi,
>>>> I have noticed that. But:
>>>> processPendingRequests is called only by enumAvailable or =
enumLeased, which are called by closeExpired and closeIdle, these are =
called explicitly right?
>>>> processNextPendingRequest is called release(explicitly called), =
requestCancelled, requestFailed, requestTimeout and these are called in =
InternalSessionRequestCallback. By corresponding methods. But as I =
understand the problem to get to this some request has to be able to do:
>>>> final SessionRequest sessionRequest =3D this.ioreactor.connect(
>>>>        remoteAddress, localAddress, route, =
this.sessionRequestCallback);
>>>> Which doesnt have to happen. And in that case the requests are =
there forever. There is no automatic remove of requests from =
leasingRequests
>>>> Tom
>>> I am sorry but I am not sure I understand the problem you are =
having.
>>> Oleg
>>>>> On 29. 4. 2015, at 11:36, Oleg Kalnichevski =
<oleg@ok2consulting.com> wrote:
>>>>> On Tue, Apr 28, 2015 at 9:14 AM, Tomas Tunkl =
<tunkl.tomas@gmail.com> wrote:
>>>>>> Hi,
>>>>>> I have some problems with ApacheAsyncHttpClient. I have tracked =
the problems to one specific place I am interested in.
>>>>>> In AbstractNIOConnPool I can see that I have in leasingRequests =
112 entries. I have searched the code and couldnt find how the entries =
are processed. I believe that request gets to leasingRequests when it =
cannot get connection, because all are used. But when the requests gets =
another chance? It seems to me, that it is stuck there forever.
>>>>> Leasing requests are processed by #processNextPendingRequest and =
#processPendingRequests methods.
>>>>> Oleg
>>>>> =
---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: httpclient-users-unsubscribe@hc.apache.org
>>>>> For additional commands, e-mail: =
httpclient-users-help@hc.apache.org
>>>> =
---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: httpclient-users-unsubscribe@hc.apache.org
>>>> For additional commands, e-mail: =
httpclient-users-help@hc.apache.org
>>> =
---------------------------------------------------------------------
>>> To unsubscribe, e-mail: httpclient-users-unsubscribe@hc.apache.org
>>> For additional commands, e-mail: httpclient-users-help@hc.apache.org
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: httpclient-users-unsubscribe@hc.apache.org =
<mailto:httpclient-users-unsubscribe@hc.apache.org>
>> For additional commands, e-mail: httpclient-users-help@hc.apache.org =
<mailto:httpclient-users-help@hc.apache.org>
>=20
>=20
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: httpclient-users-unsubscribe@hc.apache.org =
<mailto:httpclient-users-unsubscribe@hc.apache.org>
> For additional commands, e-mail: httpclient-users-help@hc.apache.org =
<mailto:httpclient-users-help@hc.apache.org>

--Apple-Mail=_CBB59347-1DD9-4F29-A6CF-EAE1E5901A9C--