<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<div class="moz-cite-prefix">El 10/04/15 a las 14:16, Alex Villacís
Lasso escribió:<br>
</div>
<blockquote cite="mid:55282195.7060505@palosanto.com" type="cite">
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
<div class="moz-cite-prefix">El 08/04/15 a las 08:22, Vinicius
Fontes escribió:<br>
</div>
<blockquote
cite="mid:CAP_GNRmwWN2UNQRdHh950wuuhUB05ZF9vOumN8Uovbr79dv73A@mail.gmail.com"
type="cite">
<div dir="ltr">Have you tried Asterisk 13? The bridging
mechanism has been completely rewritten on Asterisk 12, so
there's no longer channel masquerading and zombie channels.
Might be worth a try.</div>
<div class="gmail_extra"><br>
</div>
</blockquote>
Sorry, this client is very hard to talk into stopping its
operations long enough to install changes, let alone a major
Asterisk version change. I already had trouble convincing him of
the need to install a debugging version with DEBUG_THREADS
enabled.<br>
</blockquote>
After reviewing a "core show locks" output, I am very sure I have
hit an Asterisk bug involving a deadlock in cdr_mysql . When doing a
"core reload" on the CLI, cdr_mysql calls ast_cdr_unregister() with
its internal lock held (mysql_lock). However, writing a CDR involves
adquiring the cdr lock and then the internal lock. Therefore,
deadlock. None of the other cdr modules does that.<br>
<br>
However, I know that cdr_mysql is currently deprecated for Asterisk
11. I have a patch to fix the issue mentioned above (attached), but
I want to know how to get it reviewed, if only to see whether the
fix is sane.<br>
</body>
</html>