<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Glenn Steen wrote:
<blockquote
cite="mid:223f97700802150421q1db3beafv512da14a40308fac@mail.gmail.com"
type="cite">
<pre wrap="">On 15/02/2008, AlxFrag <a class="moz-txt-link-rfc2396E" href="mailto:alxfrag@gmail.com"><alxfrag@gmail.com></a> wrote:
</pre>
<pre wrap=""><!---->(snip)
</pre>
<blockquote type="cite">
<pre wrap=""> It says:
clamav /usr/lib/MailScanner/clamav-wrapper
/usr/local
clamd /bin/false
/usr/local
and , in /usr/lib/MailScanner/clamav-wrapper i have:
ClamScan=$1/bin/clamdscan
</pre>
</blockquote>
<pre wrap=""><!---->That is _not_ the recommended and supported way of using clamd in
MailScanner. You will pay an unnecessary fork/exec penalty for this,
compared to the very nice direct call thing Rick Cooper implemented.
The reason to not use clamavmodule (mainly, apart from some possible
build issues) and use Rick's clamd thing instead is that the
individual MS child memory footprint is decreased (using clamd)...
Since clamavmodule will have to load the signatures into every child.
Please undo those mods and look at implementing clamd the right way instead;-).
Cheers
</pre>
</blockquote>
ok thank you!<br>
</body>
</html>