<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <tt>Hi Maxim,<br>
      <br>
      Thanks a lot for the reports and your tremendous help with
      troubleshooting this. I just did the backport to 2.3.<br>
      <br>
      Best regards,<br>
    </tt>
    <pre class="moz-signature" cols="72">Bogdan-Andrei Iancu
  OpenSIPS Founder and Developer
  <a class="moz-txt-link-freetext" href="http://www.opensips-solutions.com">http://www.opensips-solutions.com</a>

OpenSIPS Bootcamp 2017, Houston, US
  <a class="moz-txt-link-freetext" href="http://opensips.org/training/OpenSIPS_Bootcamp_2017.html">http://opensips.org/training/OpenSIPS_Bootcamp_2017.html</a>
</pre>
    <div class="moz-cite-prefix">On 08/23/2017 11:19 PM, Maxim Sobolev
      wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAH7qZfuM-fvtsdXyoA9UzVQ2U6szjs0grSpRkfRJurF1ZDxtjA@mail.gmail.com">
      <div dir="ltr">Hi Bogdan, yes, that seems to be fixed it. Please
        merge into 2.3 at your earliest convenience. Thanks!
        <div><br>
        </div>
        <div>-Max</div>
      </div>
      <div class="gmail_extra"><br>
        <div class="gmail_quote">On Mon, Aug 21, 2017 at 6:29 AM,
          Bogdan-Andrei Iancu <span dir="ltr"><<a
              href="mailto:bogdan@opensips.org" target="_blank"
              moz-do-not-send="true">bogdan@opensips.org</a>></span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0
            .8ex;border-left:1px #ccc solid;padding-left:1ex">
            <div bgcolor="#FFFFFF" text="#000000"> <tt>Hi MAxim,<br>
                <br>
                Could you validate this fix:<br>
                   
                <a class="m_2422082274593172953moz-txt-link-freetext"
href="https://github.com/OpenSIPS/opensips/commit/4aa5fb91a0a8b06d886ef3f2a093c8f5526aaa4a"
                  target="_blank" moz-do-not-send="true">https://github.com/OpenSIPS/<wbr>opensips/commit/<wbr>4aa5fb91a0a8b06d886ef3f2a093c8<wbr>f5526aaa4a</a><br>
                <br>
                Regards,<br>
              </tt><span class="">
                <pre class="m_2422082274593172953moz-signature" cols="72">Bogdan-Andrei Iancu
  OpenSIPS Founder and Developer
  <a class="m_2422082274593172953moz-txt-link-freetext" href="http://www.opensips-solutions.com" target="_blank" moz-do-not-send="true">http://www.opensips-solutions.<wbr>com</a>

OpenSIPS Bootcamp 2017, Houston, US
  <a class="m_2422082274593172953moz-txt-link-freetext" href="http://opensips.org/training/OpenSIPS_Bootcamp_2017.html" target="_blank" moz-do-not-send="true">http://opensips.org/training/<wbr>OpenSIPS_Bootcamp_2017.html</a>
</pre>
              </span>
              <div>
                <div class="h5">
                  <div class="m_2422082274593172953moz-cite-prefix">On
                    08/21/2017 01:16 PM, Bogdan-Andrei Iancu wrote:<br>
                  </div>
                </div>
              </div>
              <blockquote type="cite">
                <div>
                  <div class="h5"> <tt>Hi Maxim,<br>
                      <br>
                      Sorry for the delay, finally back on tracks after
                      ClueCon,<br>
                      <br>
                      And thank you for your latest data (especially for
                      the "trap" output) - I have a good lead on what
                      might be the problem. Expect the fix soon.<br>
                      <br>
                      Best regards,<br>
                    </tt>
                    <pre class="m_2422082274593172953moz-signature" cols="72">Bogdan-Andrei Iancu
  OpenSIPS Founder and Developer
  <a class="m_2422082274593172953moz-txt-link-freetext" href="http://www.opensips-solutions.com" target="_blank" moz-do-not-send="true">http://www.opensips-solutions.<wbr>com</a>

OpenSIPS Bootcamp 2017, Houston, US
  <a class="m_2422082274593172953moz-txt-link-freetext" href="http://opensips.org/training/OpenSIPS_Bootcamp_2017.html" target="_blank" moz-do-not-send="true">http://opensips.org/training/<wbr>OpenSIPS_Bootcamp_2017.html</a>
</pre>
                    <div class="m_2422082274593172953moz-cite-prefix">On
                      08/19/2017 01:39 AM, Maxim Sobolev wrote:<br>
                    </div>
                    <blockquote type="cite">
                      <div dir="ltr">Hi Bogdan, any luck looking into
                        this issue? Looks like a serious problem with
                        2.3 / master to me. Backing out the commit in
                        question fixes it for us.
                        <div><br>
                        </div>
                        <div>-Max</div>
                        <div class="gmail_extra"><br>
                          <div class="gmail_quote">On Mon, Aug 7, 2017
                            at 12:09 PM, Maxim Sobolev <span dir="ltr"><<a
                                href="mailto:sobomax@sippysoft.com"
                                target="_blank" moz-do-not-send="true">sobomax@sippysoft.com</a>></span>
                            wrote:<br>
                            <blockquote class="gmail_quote"
                              style="margin:0 0 0 .8ex;border-left:1px
                              #ccc solid;padding-left:1ex">
                              <div dir="ltr">Bogdan, I've collected some
                                "opensipsctl trap" output here for you,
                                see line 2611 onwards:
                                <div><br>
                                </div>
                                <div><a
                                    href="https://travis-ci.org/sippy/voiptests/jobs/261951017"
                                    target="_blank"
                                    moz-do-not-send="true">https://travis-ci.org/sippy/vo<wbr>iptests/jobs/261951017</a><br>
                                </div>
                                <div><br>
                                </div>
                                <div>This is 10 seconds after starting
                                  the test, CANCELs arrive at about 2-3s
                                  mark, so it's well into "dead zone".
                                  Yet, I see some CANCEL messages in the
                                  traces, so it looks like some locking
                                  issue indeed.</div>
                                <div><br>
                                </div>
                                <div>Please let me know if it helps or
                                  if you need any more debug. For the
                                  time being, I am successfully running
                                  with 2.3 and master by reverting
                                  change 1eb4ec0f78f43f6ff546de4<wbr>9bc72e513876fb86b</div>
                                <div><br>
                                </div>
                                <div>-Max</div>
                              </div>
                              <div class="gmail_extra">
                                <div>
                                  <div class="m_2422082274593172953h5"><br>
                                    <div class="gmail_quote">On Wed, Aug
                                      2, 2017 at 8:03 AM, Bogdan-Andrei
                                      Iancu <span dir="ltr"><<a
                                          href="mailto:bogdan@opensips.org"
                                          target="_blank"
                                          moz-do-not-send="true">bogdan@opensips.org</a>></span>
                                      wrote:<br>
                                      <blockquote class="gmail_quote"
                                        style="margin:0 0 0
                                        .8ex;border-left:1px #ccc
                                        solid;padding-left:1ex">
                                        <div bgcolor="#FFFFFF"
                                          text="#000000"> <tt>Thank you
                                            Maxim for your
                                            troubleshooting.<br>
                                            <br>
                                            If you say this code makes a
                                            difference in your test, I
                                            have the say that the newly
                                            added locking may do the
                                            "blocking" . To see if this
                                            is true, while OpenSIPS is
                                            "blocked" could you run
                                            several times the
                                            "opensipsctl trap" ?<br>
                                            <br>
                                            Best regards,<br>
                                          </tt><span>
                                            <pre class="m_2422082274593172953m_7405539153921155859m_278828822372435955moz-signature" cols="72">Bogdan-Andrei Iancu
  OpenSIPS Founder and Developer
  <a class="m_2422082274593172953m_7405539153921155859m_278828822372435955moz-txt-link-freetext" href="http://www.opensips-solutions.com" target="_blank" moz-do-not-send="true">http://www.opensips-solutions.<wbr>com</a>

OpenSIPS Bootcamp 2017, Houston, US
  <a class="m_2422082274593172953m_7405539153921155859m_278828822372435955moz-txt-link-freetext" href="http://opensips.org/training/OpenSIPS_Bootcamp_2017.html" target="_blank" moz-do-not-send="true">http://opensips.org/training/O<wbr>penSIPS_Bootcamp_2017.html</a>
</pre>
                                          </span>
                                          <div>
                                            <div
                                              class="m_2422082274593172953m_7405539153921155859h5">
                                              <div
class="m_2422082274593172953m_7405539153921155859m_278828822372435955moz-cite-prefix">On
                                                08/02/2017 03:59 AM,
                                                Maxim Sobolev wrote:<br>
                                              </div>
                                              <blockquote type="cite">
                                                <div dir="ltr">Hi Bogdan
                                                  & Co, I think I've
                                                  narrowed down failure
                                                  to a specific revision
                                                  in the 2.3 branch,
                                                  see <a
                                                    href="https://travis-ci.org/sippy/voiptests/builds/260009545"
                                                    target="_blank"
                                                    moz-do-not-send="true">https://travis-ci.org/sipp<wbr>y/voiptests/builds/260009545</a>:
                                                  <div><br>
                                                  </div>
                                                  <div>
                                                    <div>commit
                                                      1eb4ec0f78f43f6ff546de49bc72e5<wbr>13876fb86b</div>
                                                    <div>Author:
                                                      Bogdan-Andrei
                                                      Iancu <<a
                                                        href="mailto:bogdan@opensips.org"
                                                        target="_blank"
moz-do-not-send="true">bogdan@opensips.org</a>></div>
                                                    <div>Date:   Thu Mar
                                                      9 15:37:01 2017
                                                      +0200</div>
                                                    <div><br>
                                                    </div>
                                                    <div>    Added
                                                      support for
                                                      dynamic branch
                                                      manipulation</div>
                                                    <div><br>
                                                    </div>
                                                    <div>    New
                                                      functions were
                                                      added to allow
                                                      remote injecting
                                                      of new branches
                                                      into an ongoing
                                                      transaction:</div>
                                                    <div>    *
                                                      t_inject_branches(source,
                                                      flags) - injects
                                                      new branches;</div>
                                                    <div>    *
                                                      t_wait_for_new_branches()
                                                      - instruct
                                                      transaction to
                                                      wait more for
                                                      possible new
                                                      injections.</div>
                                                  </div>
                                                  <div><br>
                                                  </div>
                                                  <div>Looking at the
                                                    diff in question,
                                                    the following piece
                                                    of code looks
                                                    suspicious, but it
                                                    might be something
                                                    else from that
                                                    change:<br>
                                                  </div>
                                                  <div><br>
                                                  </div>
                                                  <div>
                                                    <div>@@ -1244,7
                                                      +1301,18 @@ inline
                                                      static int
                                                      w_t_relay( struct
                                                      sip_msg  *p_msg ,
                                                      char *proxy, char
                                                      *flags)</div>
                                                    <div><br>
                                                    </div>
                                                    <div>               
update_cloned_msg_from_msg( t->uas.request, p_msg);</div>
                                                    <div><br>
                                                    </div>
                                                    <div>-              
                                                      ret =
                                                      t_forward_nonack(
                                                      t, p_msg, p);</div>
                                                    <div>+              
                                                      if
                                                      (route_type==FAILURE_ROUTE)
                                                      {</div>
                                                    <div>+              
                                                              /* If
                                                      called from
                                                      failure route we
                                                      need reset the
                                                      branch counter to</div>
                                                    <div>+              
                                                               * ignore
                                                      the previous set
                                                      of branches
                                                      (already
                                                      terminated) */</div>
                                                    <div>+              
                                                              ret =
                                                      t_forward_nonack(
                                                      t, p_msg, p,
                                                      1/*reset*/);</div>
                                                    <div>+              
                                                      } else {</div>
                                                    <div>+              
                                                              /* if
                                                      called from
                                                      request route and
                                                      the transaction
                                                      was previously</div>
                                                    <div>+              
                                                               *
                                                      created, better
                                                      lock here to avoid
                                                      any overlapping
                                                      with</div>
                                                    <div>+              
                                                               * branch
                                                      injection from
                                                      other processes */</div>
                                                    <div>+              
                                                             
                                                      LOCK_REPLIES(t);</div>
                                                    <div>+              
                                                              ret =
                                                      t_forward_nonack(
                                                      t, p_msg, p,
                                                      1/*reset*/);</div>
                                                    <div>+              
                                                             
                                                      UNLOCK_REPLIES(t);</div>
                                                    <div>+              
                                                      }</div>
                                                    <div>               
                                                      if (ret<=0 ) {</div>
                                                    <div>               
                                                             
                                                      LM_ERR("t_forward_nonack
                                                      failed\n");</div>
                                                    <div>               
                                                              ret =
                                                      t_relay_inerr2scripterr();</div>
                                                    <div>@@ -1262,7
                                                      +1330,8 @@
                                                      route_err:</div>
                                                  </div>
                                                  <div><br>
                                                  </div>
                                                  <div>The revisions
                                                    right before that
                                                    work, none after
                                                    does. Hopefully it
                                                    will give you some
                                                    clues. Let me know
                                                    if any other info is
                                                    needed.</div>
                                                  <div><br>
                                                  </div>
                                                  <div>-Max</div>
                                                </div>
                                              </blockquote>
                                              <br>
                                            </div>
                                          </div>
                                        </div>
                                      </blockquote>
                                    </div>
                                    <br>
                                    <br clear="all">
                                    <div><br>
                                    </div>
                                  </div>
                                </div>
                                <span>-- <br>
                                  <div
                                    class="m_2422082274593172953m_7405539153921155859gmail_signature"
                                    data-smartmail="gmail_signature">
                                    <div dir="ltr">Maksym Sobolyev<br>
                                      Sippy Software, Inc.<br>
                                      Internet Telephony (VoIP) Experts<br>
                                      Tel (Canada): <a
                                        href="tel:%28778%29%20783-0474"
                                        value="+17787830474"
                                        target="_blank"
                                        moz-do-not-send="true">+1-778-783-0474</a><br>
                                      Tel (Toll-Free): <a
                                        href="tel:%28855%29%20747-7779"
                                        value="+18557477779"
                                        target="_blank"
                                        moz-do-not-send="true">+1-855-747-7779</a><br>
                                      Fax: <a
                                        href="tel:%28866%29%20857-6942"
                                        value="+18668576942"
                                        target="_blank"
                                        moz-do-not-send="true">+1-866-857-6942</a><br>
                                      Web: <a
                                        href="http://www.sippysoft.com"
                                        target="_blank"
                                        moz-do-not-send="true">http://www.sippysoft.com</a><br>
                                      MSN: <a
                                        href="mailto:sales@sippysoft.com"
                                        target="_blank"
                                        moz-do-not-send="true">sales@sippysoft.com</a><br>
                                      Skype: SippySoft<br>
                                    </div>
                                  </div>
                                </span></div>
                            </blockquote>
                          </div>
                          <br>
                          <br clear="all">
                          <div><br>
                          </div>
                          -- <br>
                          <div
                            class="m_2422082274593172953gmail_signature"
                            data-smartmail="gmail_signature">
                            <div dir="ltr">Maksym Sobolyev<br>
                              Sippy Software, Inc.<br>
                              Internet Telephony (VoIP) Experts<br>
                              Tel (Canada): <a
                                href="tel:%28778%29%20783-0474"
                                value="+17787830474" target="_blank"
                                moz-do-not-send="true">+1-778-783-0474</a><br>
                              Tel (Toll-Free): <a
                                href="tel:%28855%29%20747-7779"
                                value="+18557477779" target="_blank"
                                moz-do-not-send="true">+1-855-747-7779</a><br>
                              Fax: <a href="tel:%28866%29%20857-6942"
                                value="+18668576942" target="_blank"
                                moz-do-not-send="true">+1-866-857-6942</a><br>
                              Web: <a href="http://www.sippysoft.com"
                                target="_blank" moz-do-not-send="true">http://www.sippysoft.com</a><br>
                              MSN: <a href="mailto:sales@sippysoft.com"
                                target="_blank" moz-do-not-send="true">sales@sippysoft.com</a><br>
                              Skype: SippySoft<br>
                            </div>
                          </div>
                        </div>
                      </div>
                    </blockquote>
                    <br>
                    <br>
                    <fieldset
                      class="m_2422082274593172953mimeAttachmentHeader"></fieldset>
                    <br>
                  </div>
                </div>
                <pre>______________________________<wbr>_________________
Devel mailing list
<a class="m_2422082274593172953moz-txt-link-abbreviated" href="mailto:Devel@lists.opensips.org" target="_blank" moz-do-not-send="true">Devel@lists.opensips.org</a>
<a class="m_2422082274593172953moz-txt-link-freetext" href="http://lists.opensips.org/cgi-bin/mailman/listinfo/devel" target="_blank" moz-do-not-send="true">http://lists.opensips.org/cgi-<wbr>bin/mailman/listinfo/devel</a>
</pre>
              </blockquote>
              <br>
            </div>
          </blockquote>
        </div>
        <br>
        <br clear="all">
        <div><br>
        </div>
        -- <br>
        <div class="gmail_signature" data-smartmail="gmail_signature">
          <div dir="ltr">Maksym Sobolyev<br>
            Sippy Software, Inc.<br>
            Internet Telephony (VoIP) Experts<br>
            Tel (Canada): +1-778-783-0474<br>
            Tel (Toll-Free): +1-855-747-7779<br>
            Fax: +1-866-857-6942<br>
            Web: <a href="http://www.sippysoft.com" target="_blank"
              moz-do-not-send="true">http://www.sippysoft.com</a><br>
            MSN: <a href="mailto:sales@sippysoft.com" target="_blank"
              moz-do-not-send="true">sales@sippysoft.com</a><br>
            Skype: SippySoft<br>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
  </body>
</html>