<html>
  <head>
    <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div class="moz-cite-prefix">Hello,<br>
        I have changed the configuration as below, the loss still
      appears from time to time and disappears after ipsec restart, can
      anyone help? Might there be a problem with NETKEY and IP routes?
      There isn't anything suspicious in /var/log/secure<br>
      <br>
      version 2.0     <br>
      <br>
      config setup<br>
              protostack=netkey<br>
              virtual_private=%v4:10.122.0.0/16<br>
              nhelpers=0<br>
      <br>
      conn HOST1HOST2<br>
              dpdaction=restart_by_peer<br>
              dpdtimeout=60<br>
              dpddelay=10<br>
      <br>
              left=<HOST1_public_ip><br>
              leftsourceip=10.122.7.1<br>
              leftsubnets={10.122.7.1/32}<br>
              leftid=@HOST1<br>
      <br>
              leftnexthop=%defaultroute<br>
              right=<HOST2_public_ip><br>
              rightsourceip=10.122.4.2<br>
              rightsubnets={10.122.4.0/24}<br>
              rightid=@HOST2<br>
              rightnexthop=%defaultroute<br>
      <br>
              auto=start<br>
              authby=secret<br>
              type=tunnel<br>
      <br>
      HOST1 has two interfaces, one with the public IP and the other
      with 10.122.7.1 netmask 255.255.255.255<br>
      HOST2 has two interfaces, one with the public IP and the other
      with 10.122.4.2 netmask 255.255.255.0<br>
      <br>
      net.ipv4.ip_forward = 1 on both machines. <br>
      <br>
      Both machines run CentOS 6.4 (  2.6.32-358.18.1.el6.x86_64 )<br>
      <br>
      Best regards,<br>
      Cristi<br>
      <br>
      <br>
      On 09/04/2014 12:15, Cristian Petrescu wrote:<br>
    </div>
    <blockquote cite="mid:53450FA2.2030002@telemaxvoice.ro" type="cite">
      <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
      <div class="moz-cite-prefix">Hello Nick,<br>
          Thank you for your answer. I have to keep the private IP on
        both hosts. Considering what you've told me, I will eliminate
        lo:0 and assign the 10.122.1.1 IP to a free network card on
        HOST2. I will also add left/rightsourceip. I will have to wait
        few days to see if this has solved the issue.<br>
        <br>
        Best regards,<br>
        Cristian<br>
        On 09/04/2014 11:37, Nick Howitt wrote:<br>
      </div>
      <blockquote
        cite="mid:40ab88cf44fed208118b31d9d938aba1@poweredbyclear.com"
        type="cite">
        <p>Rather than your "subnets" set up have you considered doing a
          more usual simple left/rightsubnet and adding a
          left/rightsourceip (probably 10.22.1.1 and 10.122.3.2). Then
          remove the HOST1 assignment to lo:0?</p>
        <p>Nick</p>
        <p>On 2014-04-09 09:15, Cristian Petrescu wrote:</p>
        <blockquote type="cite" style="padding-left:5px;
          border-left:#1010ff 2px solid; margin-left:5px"><!-- html ignored --><!-- head ignored --><!-- meta ignored -->
          <pre>Dear Users,
  I've been using OpenSwan CentOS 6.4 ( 2.6.32-358.2.1.el6.x86_64 ) for some time, it works well but from time to time we start experiencing around 2-3% loss on the connection. If we restart both openswan ends the loss goes away, I wasn't able to determine when and why the loss occurs. I've checked /var/log/secure and there isn't anything happening out of normal. I've had problems in the past with ksoftirqd reaching 100% but after setting /proc/sys/net/ipv4/xfrm4_gc_thresh to 100 and installing irqbalance that ksoftirqd was solved. Please help me solve this issue, below is the configuration:

HOST2:
# /etc/ipsec.conf - Openswan IPsec configuration file
#
# Manual:     ipsec.conf.5
#
# Please place your own config files in /etc/ipsec.d/ ending in .conf

version 2.0     # conforms to second version of ipsec.conf specification

# basic configuration
config setup
        protostack=netkey
        virtual_private=%v4:10.122.0.0/16
        oe=off
        nhelpers=0

conn host1host2
        dpdaction=restart_by_peer
        dpdtimeout=60
        dpddelay=10
        left=<host2 public ip>
        leftsubnets={10.122.1.1/32,<host2 public ip>/32}
        <a moz-do-not-send="true" href="mailto:leftid=@host2">leftid=@host2</a>
        leftnexthop=%defaultroute
        right=<host1 public ip>
        rightsubnets={10.122.3.0/24,<host1 public ip>/32}
        <a moz-do-not-send="true" href="mailto:rightid=@host1">rightid=@host1</a>
        rightnexthop=%defaultroute
        auto=start
        authby=secret
        type=tunnel

On HOST2, 10.122.1.1 is assigned to lo:0
On HOST1, 10.122.3.2 is assigned to a network interface that is gateway for the equipments in that network in order to reach 10.122.1.1
HOST2 and HOST1 have the same ipsec.conf configuration.

Best regards,
Cristi



_______________________________________________
<a moz-do-not-send="true" href="mailto:Users@lists.openswan.org">Users@lists.openswan.org</a>
<a moz-do-not-send="true" href="https://lists.openswan.org/mailman/listinfo/users">https://lists.openswan.org/mailman/listinfo/users</a>
Micropayments: <a moz-do-not-send="true" href="https://flattr.com/thing/38387/IPsec-for-Linux-made-easy">https://flattr.com/thing/38387/IPsec-for-Linux-made-easy</a>
Building and Integrating Virtual Private Networks with Openswan:
<a moz-do-not-send="true" href="http://www.amazon.com/gp/product/1904811256/104-3099591-2946327?n=283155">http://www.amazon.com/gp/product/1904811256/104-3099591-2946327?n=283155</a>
</pre>
        </blockquote>
      </blockquote>
      <br>
    </blockquote>
    <br>
  </body>
</html>