<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 12pt;
font-family:Calibri
}
--></style></head>
<body class='hmmessage'><div dir='ltr'>Hi Bogdan,<div><br></div><div>Thanks for the prompt response as ever!! :)</div><div><br></div><div>The goal is to move to OpenSIPS 1.11 as soon as possible.</div><div><br></div><div>We are still running 1.5, we are just upgrading the OS from Fedora, and also the Database backend, so we can do things more efficiently!</div><div><br></div><div>The next task is upgrade to 1.11 in the lab, get our config working then look to migrate.</div><div><br></div><div>Thanks for your response.</div><div><br></div><div>Jon<br><br><div><hr id="stopSpelling">Date: Wed, 11 Jun 2014 14:08:53 +0300<br>From: bogdan@opensips.org<br>To: users@lists.opensips.org; hunterj91@hotmail.com<br>Subject: Re: [OpenSIPS-Users] ERROR:usrloc:preload_udomain: fetching rows failed on OpenSIPS 1.5.3<br><br>
<div class="ecxmoz-cite-prefix"><tt>Hi Jonathan,<br>
<br>
I guess you upgraded to OpenSIPS 1.11, right ? The "fetch_rows"
parameter in usrloc was removed in 1.6 version - there is an
automatic detection of available PKG mem, and based on that, the
modules loads the proper number of rows.<br>
<br>
I understand you get that error still with 1.5, right ? if so ,
just set a lower value. The idea behind is that OpenSIPS have a
certain amount of private (PKG, per process) memory configured
(default is 1M for 1.5, if I remember correctly) ; at startup,
OpenSIPS has to load the records from location table - this
query may return a large amount of data that needs to be proxied
via PKG mem before landing into SHM mem (shared memory); and PKG
may not be enough to load the table as a whole, so you can load
it in multiple steps, fetching rows by rows.<br>
<br>
Regards,<br>
</tt>
<pre class="ecxmoz-signature">Bogdan-Andrei Iancu
OpenSIPS Founder and Developer
<a class="ecxmoz-txt-link-freetext" href="http://www.opensips-solutions.com" target="_blank">http://www.opensips-solutions.com</a></pre>
On 11.06.2014 02:21, Jonathan Hunter wrote:<br>
</div>
<blockquote cite="mid:DUB127-W3232A86151017C15A14343BD280@phx.gbl">
<style><!--
.ExternalClass .ecxhmmessage P {
padding:0px;
}
.ExternalClass body.ecxhmmessage {
font-size:12pt;
font-family:Calibri;
}
--></style>
<div dir="ltr">
<div>Hi Guys,</div>
<div><br>
</div>
<div>We are going through a process of getting in a position to
start the upgrade path from opensips 1.5.3 to the latest
release, however to achieve this, we also</div>
<div>need to move from fedora to Centos 6.5.</div>
<div><br>
</div>
<div>We are also moving from Mysql 5.1 to Percona 5.6.17.</div>
<div><br>
</div>
<div>Now having moved a working configuration from the fedora
box, over to Centos 6.5, the following error is reported when
starting opensips;</div>
<div><br>
</div>
<div><br>
</div>
<div><br>
</div>
<div>Jun 10 22:44:37 localhost /usr/local/sbin/opensips[25490]:
ERROR:core:db_allocate_rows: no memory left</div>
<div>Jun 10 22:44:37 localhost /usr/local/sbin/opensips[25490]:
ERROR:db_mysql:db_mysql_fetch_result: no memory left</div>
<div>Jun 10 22:44:37 localhost /usr/local/sbin/opensips[25490]:
ERROR:usrloc:preload_udomain: fetching rows failed</div>
<div>Jun 10 22:44:37 localhost /usr/local/sbin/opensips[25490]:
ERROR:usrloc:child_init: child(1): failed to preload domain
'location'</div>
<div>Jun 10 22:44:37 localhost /usr/local/sbin/opensips[25490]:
ERROR:core:init_mod_child: failed to initializing module
usrloc, rank 1</div>
<div>Jun 10 22:44:37 localhost /usr/local/sbin/opensips[25490]:
ERROR:core:main_loop: init_child failed for UDP listener</div>
<div>Jun 10 22:44:37 localhost /usr/local/sbin/opensips[25503]:
CRITICAL:core:receive_fd: EOF on 6</div>
<div><br>
</div>
<div><br>
</div>
<div><br>
</div>
<div>So I have added Fetch_rows entry (which allows opensips to
start and work), and I notice that the maximum I can set it to
before it crashes is 1550 rows. (was set to the default).</div>
<div><br>
</div>
<div>The server has no load, and has no memory issues
whatsoever, so we are wondering if there is a buffer issue
related to the fetch_rows() in the db connector.</div>
<div><br>
</div>
<div>Or is this simply an issue with 1.5.3 and Centos 6.5?</div>
<div><br>
</div>
<div>Any comments/thoughts would be appreciated.</div>
<div><br>
</div>
<div>As I mentioned we are pushing to make changes so we can
start moving the Opensips versions forward, its just doing it
in a controlled manner.</div>
<div><br>
</div>
<div>Many thanks</div>
<div><br>
</div>
<div>Jon</div>
<div><br>
</div>
</div>
<br>
<fieldset class="ecxmimeAttachmentHeader"></fieldset>
<br>
<pre>_______________________________________________
Users mailing list
<a class="ecxmoz-txt-link-abbreviated" href="mailto:Users@lists.opensips.org">Users@lists.opensips.org</a>
<a class="ecxmoz-txt-link-freetext" href="http://lists.opensips.org/cgi-bin/mailman/listinfo/users" target="_blank">http://lists.opensips.org/cgi-bin/mailman/listinfo/users</a>
</pre>
</blockquote>
<br></div></div> </div></body>
</html>