<!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">
On 11/15/2010 05:21 AM, Jan Zelený wrote:
<blockquote cite="mid:201011151121.12250.jzeleny@redhat.com" type="cite">
<pre wrap="">Jan Zeleny <a class="moz-txt-link-rfc2396E" href="mailto:jzeleny@redhat.com"><jzeleny@redhat.com></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Rob Crittenden <a class="moz-txt-link-rfc2396E" href="mailto:rcritten@redhat.com"><rcritten@redhat.com></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Jan Zelený wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Rob Crittenden<a class="moz-txt-link-rfc2396E" href="mailto:rcritten@redhat.com"><rcritten@redhat.com></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Jan Zelený wrote:
</pre>
<blockquote type="cite">
<pre wrap="">I tried one other solution, but this approach was recommended to me
by Pavel. It seems to be working fine. If you don't agree with the
concept (detection per request), I can present you the original one.
<a class="moz-txt-link-freetext" href="https://fedorahosted.org/freeipa/ticket/252">https://fedorahosted.org/freeipa/ticket/252</a>
Jan
</pre>
</blockquote>
<pre wrap="">
nack. I think we need some logging to say "IPA does not work with the
threaded MPM, use the pre-fork MPM" or something like that.
Otherwise it is going to silently fail and users will have no idea
why.
</pre>
</blockquote>
<pre wrap="">
I added logging as you requested. I'm still not quite sure how does the
logging work exactly, but as I understand it, this way it should be ok.
Jan
</pre>
</blockquote>
<pre wrap="">
nack, I can still run httpd.worker and serve IPA requests.
We do things to the environment so we need to be sure that each request
is isolated from all others which is why we want to run in multi-process
mode.
rob
</pre>
</blockquote>
<pre wrap="">
I based that patch on WSGI specification, which says:
"This value should evaluate true if the application object may be
simultaneously invoked by another thread in the same process, and should
evaluate false otherwise."
I didn't realize that this condition may be evaluated as false even when
running multiple threads.
That means I have to abandon this approach and try the original one. I'm
going to sync the patch I have prepared with the current HEAD and I'll
send it ASAP.
</pre>
</blockquote>
<pre wrap="">
Ok, so I've hit a small complication and I had to update the patch, so the
detection is as robust as possible. This patch is working in Fedora
environment, where the packaging of httpd is ... well ... strange at least.
The only situation it isn't handling 100% well is the case when 2 different
Apache servers are running. But since that's not common case and I can't think
of a way to detect which instance is running current WSGI script, the script
just doesn't allow such situation.
Jan
</pre>
<pre wrap="">
<fieldset class="mimeAttachmentHeader"></fieldset>
_______________________________________________
Freeipa-devel mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Freeipa-devel@redhat.com">Freeipa-devel@redhat.com</a>
<a class="moz-txt-link-freetext" href="https://www.redhat.com/mailman/listinfo/freeipa-devel">https://www.redhat.com/mailman/listinfo/freeipa-devel</a></pre>
</blockquote>
ACK and pushed to master<br>
</body>
</html>