RE: mod_python import failure on FreeBSD AMD64
Surly The Clown <[email protected]> Tue, 4 May 2010 20:13:51 -0700
| Newsgroups | gmane.comp.python.mod_python |
|---|---|
| Message-ID | <[email protected]> |
--===============1791138638==
Content-Type: multipart/alternative;
boundary="_9ea15508-6b2e-4c24-a847-788ba1434642_"
--_9ea15508-6b2e-4c24-a847-788ba1434642_
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
> It isn't actively developed any more and if all you want to do is host
> a web application=2C and not hook into Apache request phases=2C you are
> better off using WSGI capable Python web frameworks and hosting using
> mod_wsgi or fastcgi/flup.
Unfortunately=2C I do need to hook into Apache request phases...
>=20
> > I tried setting PythonOption mod_python.mutex_locks 4=2C and the proble=
m
> > persisted. I then tried increasing the number of available semaphores i=
n the
> > kernel from the default 10 (!) to 256=2C verified that that setting had
> > changed=2C and then tried again. The problem still persists. Furthermor=
e=2C four
> > files that appear to be the mutexes exist in /tmp=2C
>=20
> The sem_init() call is I believe a kernel thing=2C not a filesystem type =
of lock.
>=20
> Such kernel setting changes usually only take affect after a system
> reboot. Did you do that?
Sure did.
>=20
> > named mpmtx15570 to
> > mpmtx15573. Perhaps mod_python isn't using kernel semaphores? Its
> > documentation says that it doesn't always. And /tmp isn't out of space =
(no
> > partition is).
>=20
> What is used depends on what Apache/APR sets up as default type of
> semaphore. It prefers certain types and what is available depends on
> the operating system.
I might have to dig into that to see where the sem_init message is coming f=
rom.
>=20
> > Any more ideas?
>=20
> Don't use mod_python if you don't have to.
>=20
> What are you wanting to host using it?
Big(ish) media files=2C and I'm wanting to hook into Apache so that I can c=
ontrol access to them with custom logic. As far as I've been able to determ=
ine=2C this is something that wsgi/fastcgi can't really do=2C unless you us=
e URL rewriting to force the requests to go through a script. Even then=2C =
it's suboptimal as the file would then be piped through a script instead of=
handled directly by Apache. mod_python was pretty much exactly what I need=
.
Thank you for your attempts to help=2C at any rate.
=20
_________________________________________________________________
Win a $10=2C000 shopping spree from Hotmail! Enter now.
http://go.microsoft.com/?linkid=3D9729711=
--_9ea15508-6b2e-4c24-a847-788ba1434642_
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<html>
<head>
<style><!--
.hmmessage P
{
margin:0px=3B
padding:0px
}
body.hmmessage
{
font-size: 10pt=3B
font-family:Verdana
}
--></style>
</head>
<body class=3D'hmmessage'>
>=3B It isn't actively developed any more and if all you want to do is ho=
st<BR>>=3B a web application=2C and not hook into Apache request phases=
=2C you are<BR>>=3B better off using WSGI capable Python web frameworks a=
nd hosting using<BR>>=3B mod_wsgi or fastcgi/flup.<BR><BR><BR>Unfortunate=
ly=2C I do need to hook into Apache request phases...<BR><BR><BR>>=3B <BR=
>>=3B >=3B I tried setting PythonOption mod_python.mutex_locks 4=2C and=
the problem<BR>>=3B >=3B persisted. I then tried increasing the number=
of available semaphores in the<BR>>=3B >=3B kernel from the default 10=
(!) to 256=2C verified that that setting had<BR>>=3B >=3B changed=2C a=
nd then tried again. The problem still persists. Furthermore=2C four<BR>>=
=3B >=3B files that appear to be the mutexes exist in /tmp=2C<BR>>=3B <=
BR>>=3B The sem_init() call is I believe a kernel thing=2C not a filesyst=
em type of lock.<BR>>=3B <BR>>=3B Such kernel setting changes usually o=
nly take affect after a system<BR>>=3B reboot. Did you do that?<BR><BR><B=
R>Sure did.<BR><BR><BR>>=3B <BR>>=3B >=3B named mpmtx15570 to<BR>>=
=3B >=3B mpmtx15573. Perhaps mod_python isn't using kernel semaphores? It=
s<BR>>=3B >=3B documentation says that it doesn't always. And /tmp isn'=
t out of space (no<BR>>=3B >=3B partition is).<BR>>=3B <BR>>=3B Wha=
t is used depends on what Apache/APR sets up as default type of<BR>>=3B s=
emaphore. It prefers certain types and what is available depends on<BR>>=
=3B the operating system.<BR><BR><BR>I might have to dig into that to see w=
here the sem_init message is coming from.<BR><BR><BR>>=3B <BR>>=3B >=
=3B Any more ideas?<BR>>=3B <BR>>=3B Don't use mod_python if you don't =
have to.<BR>>=3B <BR>>=3B What are you wanting to host using it?<BR><BR=
><BR>Big(ish) media files=2C and I'm wanting to hook into Apache so that I =
can control access to them with custom logic. As far as I've been able to d=
etermine=2C this is something that wsgi/fastcgi can't really do=2C unless y=
ou use URL rewriting to force the requests to go through a script. Even the=
n=2C it's suboptimal as the file would then be piped through a script inste=
ad of handled directly by Apache. mod_python was pretty much exactly what I=
need.<BR><BR><BR>Thank you for your attempts to help=2C at any rate.<BR><B=
R><BR> <br /><hr />Win a $10=2C000 shopping spree from Hotmail! =
Enter now <a href=3D'http://go.microsoft.com/?linkid=3D9729706' target=3D'_=
new'>Enter now</a></body>
</html>=
--_9ea15508-6b2e-4c24-a847-788ba1434642_--
--===============1791138638==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
_______________________________________________
Mod_python mailing list
[email protected]
http://mailman.modpython.org/mailman/listinfo/mod_python
--===============1791138638==--