[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFS: streamline



Hi Jurij,

On Wednesday 01 September 2004 12:48 am, Jurij Smakov wrote:
> I have a somewhat different impression, apache2 seems to have its share
> of problems. Debian's apache2 packages have seen 12 uploads during the
> last two months [0] 

Yes, there was a botched and now reverted attempt at enabling LFS, but that's 
not really relevant here.  =)

Other than that, I have had an excellent with apache2 so far.

> and PHP manual explicitly states [1]: 
>
>  Do not use Apache 2.0.x and PHP in a production environment neither on
>  Unix nor on Windows. For information on why, read the following FAQ
>  entry.

That FAQ entry just says that PHP isn't threadsafe.  There are plenty of other 
reasons to use Apache 2 other than the threaded MPMs, and PHP works just fine 
with the prefork MPM.  I think that FAQ entry is somewhat shortsighted and 
even a bit FUD-ful.

> While seemingly innocent, this change may have some "interesting"
> consequences. Consider a situation when apache2 is provided by the
> recommended apache2-mpm-worker package. If PHP is not installed, the first
> candidate to satisfy the second dependency is libapache2-mod-php4, which
> depends on apache2-mpm-prefork (>= 2.0.50-10). 

> This, in turn, conflicts 
> with all other apache2 MPM flavors (including apache2-mpm-worker). As a
> result, the apache2-mpm-worker is going to be (quite unexpectedly) removed
> and replaced by apache2-mpm-prefork, possibly, contrary to user's wishes.

No, it would not be removed unexpectedly -- the user would have to say OK 
after being told very clearly that apache2-mpm-worker was to be removed.  If 
a user has that MPM installed, this should be a clear indicator to them that 
they should not proceed with the installation of this package, as it is 
simply not compatible.  The only way they could use Apache 2 with your 
package is to switch to the prefork MPM, which is exactly what APT would 
offer to do.  =)

Cheers,
nate

-- 
Nathaniel W. Turner
http://houseofnate.net/



Reply to: