Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 27 Oct 2008 23:47:12 -0300
From:      Jose Amengual <jose.amengual@gmail.com>
To:        Bjoern A. Zeeb <bzeeb-lists@lists.zabbadoz.net>
Cc:        freebsd-current@freebsd.org, freebsd-stable@freebsd.org
Subject:   SOLUTION!!!! Re: tcp_do_segment....again
Message-ID:  <D983434B-3A49-4378-B56E-376DDFC901E9@gmail.com>
In-Reply-To: <20081027120802.G2978@maildrop.int.zabbadoz.net>
References:  <D7875010-7269-4CDF-99CA-6357230B4CB5@gmail.com> <20081027120802.G2978@maildrop.int.zabbadoz.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Hi Guys.

I will describe de steps that I did to solve this problem that I still  
don't understand.

Like my first post, I installed FreeBSD with the install cd 7.0- 
STABLE-200807-i386-disc1.iso , then I update using cvsup and the  
stable supfile and I got the 7.1-PREREALEASE ( I don't know why I got  
7.1 has a stable using the stablesupfile using RELENG_7 ).

After make world and stuff, I installed ezjail and I start creating my  
jails, then I configure my pf script and I give my jails and user nat  
and etc.
Then I start installing packages to my jails and I start seeing the  
TCP kernel messages that at the first time I ignored.

After I had all the jails working , I realize that my postfix jail was  
leaving almost all the emails in the defer queue with reasons like  
"Connection time out", then I realize that the tcp_do_segment messages  
were talking about port 25 connections and not only ftp connections,  
and that was exactly one message per mail in the queue.

After reading some post I decide to Update my machine again to  
RELENG_7_0 and I got 7.0-p5 and the problem persist, after that I  
start changing sysctl variables like sack etc.

Today I reinstall the same machine with the 7.0-RELEASE-i386-disc1.iso  
and update to 7.0-p5 and using ezjail I start creating the jails and I  
realize that I was getting the same TCP messages per ftp connections  
but seems to be working ok. After some research and reading I decide  
to REMOVE the SCTP option from the kernel and recompile.

After that I didn't receive any message and today the server is up and  
running with 300 users behind and tons of email with NONE in the defer  
queue.

The configurations of the jails were copy and paste to the new  
installation of the server, the only thing that I change was the  
install CD.

Is that possible ?, could be the reason ?
SCTP is generating this problem ?



Thanks to all.


On 27-Oct-08, at 9:09 AM, Bjoern A. Zeeb wrote:

> On Sun, 26 Oct 2008, Jose Amengual wrote:
>
> Hi,
>
>> The problem persist even in stable 7.0, 7.1-PRERELEASE and 7.0-P5 ,  
>> the postfix jail is living all the email in defer queue saying  
>> "Connection Time Out"
>>
>> This server give service to 300 clients and the network load os high.
>>
>> is there any clue how to resolve this ?
>
> sorry if I am not reading back various links which include mails I had
> sent which are entirely unrelated.
>
> So you are seeing TCP problems. And you are running with a firewall
> enabled?
>
> Have you tried tcpdumping and seeing what exactly goes wrong?
>
> do you have any offload stuff enabled (as the subject suggests)? If so
> what happens if you disable this?
>
> /bz
>
> -- 
> Bjoern A. Zeeb              Stop bit received. Insert coin for new  
> game.
> _______________________________________________
> freebsd-current@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to "freebsd-current-unsubscribe@freebsd.org 
> "




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?D983434B-3A49-4378-B56E-376DDFC901E9>