apache/fcgi deprecation?

apache/fcgi deprecation?

[Nathan Cutler]

I was just browsing the RGW documentation and noticed that parts of it 
are still apache/fcgi-oriented (with no mention of civetweb) even in master.

My impression was that civetweb is now recommended for RGW deployments 
as it is faster. Is that true? And is there a plan to deprecate/drop 
apache/fcgi support, or will it remain indefinitely?

Would it be a good idea to expunge apache/fcgi from the docs and point 
folks to civetweb instead?

-- 
Nathan Cutler
Software Engineer Distributed Storage
SUSE LINUX, s.r.o.
Tel.: +420 284 084 037

Re: apache/fcgi deprecation?

[Ken Dreyer]

Sure, please do it!

I think it's just a matter of no one having time to remove it.

On Thu, Feb 4, 2016 at 9:12 PM, Nathan Cutler <ncutler@suse.cz> wrote:
> I was just browsing the RGW documentation and noticed that parts of it are
> still apache/fcgi-oriented (with no mention of civetweb) even in master.
>
> My impression was that civetweb is now recommended for RGW deployments as it
> is faster. Is that true? And is there a plan to deprecate/drop apache/fcgi
> support, or will it remain indefinitely?
>
> Would it be a good idea to expunge apache/fcgi from the docs and point folks
> to civetweb instead?
>
> --
> Nathan Cutler
> Software Engineer Distributed Storage
> SUSE LINUX, s.r.o.
> Tel.: +420 284 084 037
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: apache/fcgi deprecation?

[Nathan Cutler]

On 02/08/2016 11:57 PM, Ken Dreyer wrote:
> Sure, please do it!
>
> I think it's just a matter of no one having time to remove it.

So we could remove the FastCGI dependency from the packaging, too?

> On Thu, Feb 4, 2016 at 9:12 PM, Nathan Cutler <ncutler@suse.cz> wrote:
>> I was just browsing the RGW documentation and noticed that parts of it are
>> still apache/fcgi-oriented (with no mention of civetweb) even in master.
>>
>> My impression was that civetweb is now recommended for RGW deployments as it
>> is faster. Is that true? And is there a plan to deprecate/drop apache/fcgi
>> support, or will it remain indefinitely?
>>
>> Would it be a good idea to expunge apache/fcgi from the docs and point folks
>> to civetweb instead?
>>
>> --
>> Nathan Cutler
>> Software Engineer Distributed Storage
>> SUSE LINUX, s.r.o.
>> Tel.: +420 284 084 037
>> --
>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html


-- 
Nathan Cutler
Software Engineer Distributed Storage
SUSE LINUX, s.r.o.
Tel.: +420 284 084 037

Re: apache/fcgi deprecation?

[Sage Weil]

Matt, Yehuda-

What do you think?  There are some valid use-cases for wanting apache (or 
whatever) in front as it gives you access to a bunch of modules.  On the 
other hand, you can do the same with a proxy.

sage


On Tue, 9 Feb 2016, Nathan Cutler wrote:

> On 02/08/2016 11:57 PM, Ken Dreyer wrote:
> > Sure, please do it!
> > 
> > I think it's just a matter of no one having time to remove it.
> 
> So we could remove the FastCGI dependency from the packaging, too?
> 
> > On Thu, Feb 4, 2016 at 9:12 PM, Nathan Cutler <ncutler@suse.cz> wrote:
> > > I was just browsing the RGW documentation and noticed that parts of it are
> > > still apache/fcgi-oriented (with no mention of civetweb) even in master.
> > > 
> > > My impression was that civetweb is now recommended for RGW deployments as
> > > it
> > > is faster. Is that true? And is there a plan to deprecate/drop apache/fcgi
> > > support, or will it remain indefinitely?
> > > 
> > > Would it be a good idea to expunge apache/fcgi from the docs and point
> > > folks
> > > to civetweb instead?
> > > 
> > > --
> > > Nathan Cutler
> > > Software Engineer Distributed Storage
> > > SUSE LINUX, s.r.o.
> > > Tel.: +420 284 084 037
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> > > the body of a message to majordomo@vger.kernel.org
> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 
> -- 
> Nathan Cutler
> Software Engineer Distributed Storage
> SUSE LINUX, s.r.o.
> Tel.: +420 284 084 037
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> 

Re: apache/fcgi deprecation?

[Yehuda Sadeh-Weinraub]

Right, apache can be used as a proxy. In any case, my understanding is
that it's not a dependency, but we'll still have the ability to
install it (mod-proxy-fcgi) if needed (correct me if I'm wrong here).
There aren't any current plans that I know of to remove the rgw fcgi
frontend.

Yehuda


On Mon, Feb 8, 2016 at 4:30 PM, Sage Weil <sweil@redhat.com> wrote:
> Matt, Yehuda-
>
> What do you think?  There are some valid use-cases for wanting apache (or
> whatever) in front as it gives you access to a bunch of modules.  On the
> other hand, you can do the same with a proxy.
>
> sage
>
>
> On Tue, 9 Feb 2016, Nathan Cutler wrote:
>
>> On 02/08/2016 11:57 PM, Ken Dreyer wrote:
>> > Sure, please do it!
>> >
>> > I think it's just a matter of no one having time to remove it.
>>
>> So we could remove the FastCGI dependency from the packaging, too?
>>
>> > On Thu, Feb 4, 2016 at 9:12 PM, Nathan Cutler <ncutler@suse.cz> wrote:
>> > > I was just browsing the RGW documentation and noticed that parts of it are
>> > > still apache/fcgi-oriented (with no mention of civetweb) even in master.
>> > >
>> > > My impression was that civetweb is now recommended for RGW deployments as
>> > > it
>> > > is faster. Is that true? And is there a plan to deprecate/drop apache/fcgi
>> > > support, or will it remain indefinitely?
>> > >
>> > > Would it be a good idea to expunge apache/fcgi from the docs and point
>> > > folks
>> > > to civetweb instead?
>> > >
>> > > --
>> > > Nathan Cutler
>> > > Software Engineer Distributed Storage
>> > > SUSE LINUX, s.r.o.
>> > > Tel.: +420 284 084 037
>> > > --
>> > > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> > > the body of a message to majordomo@vger.kernel.org
>> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>
>> --
>> Nathan Cutler
>> Software Engineer Distributed Storage
>> SUSE LINUX, s.r.o.
>> Tel.: +420 284 084 037
>> --
>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
>>
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: apache/fcgi deprecation?

[Matt Benjamin]

Hi,

I think that is correct, at least in RHEL.  As long as we continue to support the fcgi integration in rgw, this should work.

There really does seem to be a lot of application of various header, logging, and cache-control customization setups that we don't want to emulate in rgw itself.  I'm interested in hearing feedback from folks about the acceptability/preference for or against proxy vs. embedding, esp. larger deployments.

Matt

----- Original Message -----
> From: "Yehuda Sadeh-Weinraub" <yehuda@redhat.com>
> To: "Sage Weil" <sweil@redhat.com>
> Cc: "Nathan Cutler" <ncutler@suse.cz>, "Ken Dreyer" <kdreyer@redhat.com>, "Ceph Development"
> <ceph-devel@vger.kernel.org>
> Sent: Monday, February 8, 2016 8:16:05 PM
> Subject: Re: apache/fcgi deprecation?
> 
> Right, apache can be used as a proxy. In any case, my understanding is
> that it's not a dependency, but we'll still have the ability to
> install it (mod-proxy-fcgi) if needed (correct me if I'm wrong here).
> There aren't any current plans that I know of to remove the rgw fcgi
> frontend.
> 
> Yehuda
> 
> 
> On Mon, Feb 8, 2016 at 4:30 PM, Sage Weil <sweil@redhat.com> wrote:
> > Matt, Yehuda-
> >
> > What do you think?  There are some valid use-cases for wanting apache (or
> > whatever) in front as it gives you access to a bunch of modules.  On the
> > other hand, you can do the same with a proxy.
> >
> > sage
> >
> >
> > On Tue, 9 Feb 2016, Nathan Cutler wrote:
> >
> >> On 02/08/2016 11:57 PM, Ken Dreyer wrote:
> >> > Sure, please do it!
> >> >
> >> > I think it's just a matter of no one having time to remove it.
> >>
> >> So we could remove the FastCGI dependency from the packaging, too?
> >>
> >> > On Thu, Feb 4, 2016 at 9:12 PM, Nathan Cutler <ncutler@suse.cz> wrote:
> >> > > I was just browsing the RGW documentation and noticed that parts of it
> >> > > are
> >> > > still apache/fcgi-oriented (with no mention of civetweb) even in
> >> > > master.
> >> > >
> >> > > My impression was that civetweb is now recommended for RGW deployments
> >> > > as
> >> > > it
> >> > > is faster. Is that true? And is there a plan to deprecate/drop
> >> > > apache/fcgi
> >> > > support, or will it remain indefinitely?
> >> > >
> >> > > Would it be a good idea to expunge apache/fcgi from the docs and point
> >> > > folks
> >> > > to civetweb instead?
> >> > >
> >> > > --
> >> > > Nathan Cutler
> >> > > Software Engineer Distributed Storage
> >> > > SUSE LINUX, s.r.o.
> >> > > Tel.: +420 284 084 037
> >> > > --
> >> > > To unsubscribe from this list: send the line "unsubscribe ceph-devel"
> >> > > in
> >> > > the body of a message to majordomo@vger.kernel.org
> >> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>
> >>
> >> --
> >> Nathan Cutler
> >> Software Engineer Distributed Storage
> >> SUSE LINUX, s.r.o.
> >> Tel.: +420 284 084 037
> >> --
> >> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> >> the body of a message to majordomo@vger.kernel.org
> >> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >>
> >>
> > --
> > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

-- 
-- 
Matt Benjamin
Red Hat, Inc.
315 West Huron Street, Suite 140A
Ann Arbor, Michigan 48103

http://www.redhat.com/en/technologies/storage

tel.  734-707-0660
fax.  734-769-8938
cel.  734-216-5309

Re: apache/fcgi deprecation?

[Blair Bethwaite]

On 9 February 2016 at 11:30, Sage Weil <sweil@redhat.com> wrote:
> What do you think?  There are some valid use-cases for wanting apache (or
> whatever) in front as it gives you access to a bunch of modules.  On the
> other hand, you can do the same with a proxy.

I don't think it's valid to point people at the civetweb docs without
some clear info about how it is embedded in radosgw and how (if it's
even possible?) to change the embedded civetweb's config.

At Monash we recently implemented radosgw under Red Hat Ceph (i.e.,
Hammer .3-ish). We were quite surprised (and to be honest, a bit
disappointed) that there was apparently no documentation about
architecture or configuration for standing up a production grade
radosgw service with the current tech (civetweb) - understandable to
see upstream/community docs lacking, but had hoped the Red Hat
subscription might unlock some useful internal docs/configs to this
effect.

After some effort we now have a horizontally scalable radosgw behind
haproxy (SSL termination there) + keepalived with basic DNS
round-robin providing the initial request fan-out, and would be
willing to contribute some rudimentary info to the docs around this.
We're now trying to figure out how we can implement some basic
security controls with this setup (e.g., block brute-force
attempts)...

-- 
Cheers,
~Blairo

Re: apache/fcgi deprecation?

[Ken Dreyer]

On Mon, Feb 8, 2016 at 4:07 PM, Nathan Cutler <ncutler@suse.cz> wrote:
> On 02/08/2016 11:57 PM, Ken Dreyer wrote:
>>
>> Sure, please do it!
>>
>> I think it's just a matter of no one having time to remove it.
>
>
> So we could remove the FastCGI dependency from the packaging, too?
>

That would be nice, but currently RGW hard-codes a dependency on
fcgi-devel, and there's no "--without-fcgi" option to ./configure

- Ken