rabbit-proxy-users — Normal usage and help

Re: [Rabbit-proxy-users] Failure to log in...

[Dave <tec...@ml...>]

----- Original Message ----- 
From: "Samuel Hill" <Sam...@Co...>
To: <rab...@li...>
Sent: Wednesday, June 23, 2004 7:00 AM
Subject: [Rabbit-proxy-users] Failure to log in...


> When going to the following page it is supposed to ask for a login but
> does not prompt for one.
> http://206.159.167.148/jbcintranet/
>
> Avoiding the proxy works fine.
>
> Sam
>


If you are trying to access the management screens, you need to make sure
that the DNS entries match the hostname for the computer, and that the
browser you are using is set to use the proxy.

Dave

[Rabbit-proxy-users] Failure to log in...

[Samuel H. <Sam...@Co...>]

When going to the following page it is supposed to ask for a login but
does not prompt for one.
http://206.159.167.148/jbcintranet/

Avoiding the proxy works fine.

Sam

[Rabbit-proxy-users] 2.0.32 released

[Robert O. <ro...@kh...>]

Hello!

I have released the 2.0.32 now.
Report bugs and suggestions for the next version, please.

You can download it from:
http://www.khelekore.org/rabbit/

Changelog avaiable at:
http://www.khelekore.org/rabbit/changelog.shtml

Have fun
/robo

[Rabbit-proxy-users] Beta of 2.0.32

[Robert O. <ro...@kh...>]

Hello!

As I said before, I plan to release the next release soon.

There is a beta of 2.0.32 available from
http://www.khelekore.org/rabbit/RabbIT2.0.32-pre8.tar.gz

Any help with final testing is appreciated.

The 2.0.32 release is mostly memory optimization so that rabbit
can handle big caches. There are a few other fixes in there as
well like: HTTP/0.9 handling, ssl proxy chaining, rejected client
handling and admin page handling.

Have fun.
/robo

[Rabbit-proxy-users] about the site

[Robert O. <ro...@di...>]

Hello.

The khelekore.org site is currently down. It will be back up soon
(hopefully back up on thursday).The adsl modem seems to
have gone kaputt.

The sf.net site still works and I read the mailing list, but a bit less
often (since I currently only have network access at work).

Rabbit proxy is still in development, the next release will probably be
shortly after the site is up again.

Sam: do you have any news about the pre6 version?

/robo

RE: [Rabbit-proxy-users] Site/Files

[Samuel H. <Sam...@Co...>]

Maybe it is just his server is down.

I could offer hosting if I could get a hold of him.
The mail and web server are the same one.

Sam


-----Original Message-----
From: rab...@li...
[mailto:rab...@li...] On Behalf Of Tom
Grove
Sent: Friday, June 04, 2004 9:16 AM
To: rab...@li...
Subject: Re: [Rabbit-proxy-users] Site/Files


I haven't been able to get to the site either.  This is not good because
as you, I am using it in a production environment.  If he is reading
these and has plans of stopping productions please give someone the okay
to continue development.  Thanks.

-TG

========================================
>I have been unable to reach via email or by the web page the author and

>files for this project. This is a week now.
>
>Is it just me?
>
>Sam



-------------------------------------------------------
This SF.Net email is sponsored by the new InstallShield X.
From Windows to Linux, servers to mobile, InstallShield X is the one
installation-authoring solution that does it all. Learn more and
evaluate today! http://www.installshield.com/Dev2Dev/0504
_______________________________________________
Rabbit-proxy-users mailing list Rab...@li...
https://lists.sourceforge.net/lists/listinfo/rabbit-proxy-users

Re: [Rabbit-proxy-users] Site/Files

[Tom G. <tg...@wi...>]

I haven't been able to get to the site either.  This is not good because as you, I am using it in a production environment.  If he is reading these and has plans of stopping productions please give someone the okay to continue development.  Thanks.

-TG

========================================
>I have been unable to reach via email or by the web page the author and
>files for this project.
>This is a week now.
>
>Is it just me?
>
>Sam

[Rabbit-proxy-users] Site/Files

[Samuel H. <Sam...@Co...>]

I have been unable to reach via email or by the web page the author and
files for this project.
This is a week now.

Is it just me?

Sam
 

Re: [Rabbit-proxy-users] HTTPS

[Costa T. <kt...@ho...>]

> Ok, it seems that proxy chaning and ssl did not work
> well in rabbit, I have created a simple fix.
> 
> http://www.khelekore.org/rabbit/RabbIT2.0.32-pre5.tar.gz
> 
> Please help test it.
> It works fine for me with two rabbit proxies in a chain.
> 
> If it fails for you can you give me the network traffic
> that rabbit sends and recieves? (using ethereal or something
> similar).

It seems to work fine now.

Thanks!

Costa

Re: [Rabbit-proxy-users] HTTPS

[Robert O. <ro...@kh...>]

Costa Tsaousis wrote:

>Since you mentioned "my setup", I tested HTTPS with and without a back-end
>proxy (proxyhost, proxyport). It seems that the problem appears only when
>RabbIT2 is configured to use another proxy . I verified that if RabbIT2
>is standalone HTTPS is working properly.
>  
>
Ok, it seems that proxy chaning and ssl did not work
well in rabbit, I have created a simple fix.

http://www.khelekore.org/rabbit/RabbIT2.0.32-pre5.tar.gz

Please help test it.
It works fine for me with two rabbit proxies in a chain.

If it fails for you can you give me the network traffic
that rabbit sends and recieves? (using ethereal or something
similar).

/robo

Re: [Rabbit-proxy-users] HTTPS

[Robert O. <ro...@kh...>]

Costa Tsaousis wrote:

>I am testing RabbIT2 for providing my users with an Internet Accelerator.
>I have about 10.000 modems to serve 
>
Ok, bigger than I expected.

>It is not necessary to chain proxy SSL, but how
>can I tell this to RabbIT2?
>  
>
Currently you can not, but you tell the clients to
use another proxy for ssl. Having a proxy chain for ssl
introduce latency and that is one of the things you want
to minimize for a proxy accellerator (that is why you
have keepalive/persistent connections in HTTP/1.1).
Since none of the proxies in the chain can cache the
content it does not make much sense to have a chain for ssl.
Client setup instructions may be simpler and your firewall
setup may also be easier.

I will look more at rabbit and see if I can fix the
problem you have, so hopefully you wont have to
tell rabbit anything.

> <>Up to now, I plan the whole setup as follows: ...
> What do you think?

Well your setup makes sense. It is a bigger setup
than I expected and for big sites you want
redundancy and failover. So having squid for
caching and load balancing seems wise.
Having 1.5 TB of cache for rabbit is also something
I would recomend against (rabbit was not designed for
that size it was designed to handle a couple of GB,
may work for a bit more, but that is untested). So having
the netcaches behind rabbit seems like a good idea.

And to be honest, I have not handled that many clients
myself with any proxy so you are probably better qualified
to design the setup.

Have fun
/robo

Re: [Rabbit-proxy-users] HTTPS

[Costa T. <kt...@ho...>]

> What proxies are you trying to chain? and why?

I am testing RabbIT2 for providing my users with an Internet Accelerator.
I have about 10.000 modems to serve (not all of them will use
acceleration, but many will do - it will be a free service). There is a
large cluster of Network Appliances "NetCache C6XXX" (proxies) as the
core proxies for the network. I should use these for all outgoing
connections, since most probably the objects RabbIT2 will try to fetch
will already be there. It is not necessary to chain proxy SSL, but how
can I tell this to RabbIT2?

Up to now, I plan the whole setup as follows:

a. The back-end proxies will be the NetCaches since these are highly
optimized caches and they have a huge cache (1,5TB).

b. In front of these, there will be an array of RabbIT2 machines in
compression only mode (no cache).

c. In front of RabbIT2, there will be a single cluster of squid servers
configured to load balance the RabbIT2 servers, that will provide the
cache.

I plan to do it this way because:

1. I trust squid for its reliability. In this setup, if a RabbIT2 server
fails, squid is smart enough to re-route its requests to another RabbIT2
server.

2. Squid supports ICP for inter-squid communication about the objects in
the cache and the objects being fetched. This will ensure that I will be
compressing each object once.

3. I will be able to add as many RabbIT2 machines as needed without
worring about load balancing them. Squid will do it.

What do you think?

Costa

Re: [Rabbit-proxy-users] HTTPS

[Robert O. <ro...@kh...>]

Costa Tsaousis wrote:

>Since you mentioned "my setup", I tested HTTPS with and without a back-end
>proxy (proxyhost, proxyport). It seems that the problem appears only when
>RabbIT2 is configured to use another proxy . I verified that if RabbIT2
>is standalone HTTPS is working properly.
>  
>
Ok, so rabbit fails to proxy chain ssl. I will see if
I can verify this and if I can it ought to be easy to fix.
However why do you have to proxy chain ssl? None of the
proxies can read the data so all proxies will only tunnel
the connection.

What proxies are you trying to chain? and why?

I would guess that you have a squid proxy. Note that rabbit
is better at HTTP/1.1 than squid so I would expect rabbit to
handle the cache _slightly_ better (squid is a _good_ proxy).

/robo

Re: [Rabbit-proxy-users] HTTPS

[Costa T. <kt...@ho...>]

> Rabbit can proxy https request, but it can not
> filter the streams (since that requires cracking the
> encryption).

Yes, I know.
 
> >Now RabbIT2 logs a CONNECT but the browser gives a DNS error.
> >With other proxies, or directly, the URLs tried are working.
> >  
> >
> I am not sure I understand what error you get
> here. Last time I checked both https proxying and
> proxy chaining worked correctly. It would be
> easier if you told me more about the setup
> you are using.

For this test I am just using a RabbIT 2.0.32-pre4 as proxy. Lets say IE6
is trying to fetch:

https://digitalid.verisign.com/

When IE6 is attempting to load this, I see in RabbIT2 logs:

X.X.X.X - - 22/May/2004:22:16:02 GMT "CONNECT digitalid.verisign.com:443
HTTP/1.0" 200 - 

but instead of delivering the page, IE6 just displays:

---

The page cannot be displayed 
The page you are looking for is currently unavailable. The Web site might
be experiencing technical difficulties, or you may need to adjust your
browser settings. 

... blah blah...

Cannot find server or DNS Error
Internet Explorer

---

In my config there is:

allowSSL=443,444,445

Since you mentioned "my setup", I tested HTTPS with and without a back-end
proxy (proxyhost, proxyport). It seems that the problem appears only when
RabbIT2 is configured to use another proxy . I verified that if RabbIT2
is standalone HTTPS is working properly.

Costa

Re: [Rabbit-proxy-users] HTTPS

[Robert O. <ro...@kh...>]

Costa Tsaousis wrote:

>Is there something to be configured to make HTTPS connections work through
>RabbIT2? I have declared port 443 as HTTPS.
>  
>
Rabbit can proxy https request, but it can not
filter the streams (since that requires cracking the
encryption).

>Now RabbIT2 logs a CONNECT but the browser gives a DNS error.
>With other proxies, or directly, the URLs tried are working.
>  
>
I am not sure I understand what error you get
here. Last time I checked both https proxying and
proxy chaining worked correctly. It would be
easier if you told me more about the setup
you are using.

/robo

[Rabbit-proxy-users] HTTPS

[Costa T. <kt...@ho...>]

Is there something to be configured to make HTTPS connections work through
RabbIT2? I have declared port 443 as HTTPS.

Now RabbIT2 logs a CONNECT but the browser gives a DNS error.
With other proxies, or directly, the URLs tried are working.

Costa

Re: [Rabbit-proxy-users] How to compress but not cache?

[Robert O. <ro...@kh...>]

Costa Tsaousis wrote:

>I am going to have RabbIT2 between two other HTTP caches. How can I make
>it compress but not cache anything at all?
>
>When I use the DontCacheFilter it seems that it does not convert images
>(although it compresses HTMLs).
>  
>
To recode images rabbit need a cache. You can set the
cache size to 0 though.

A 0-sized cache will still create files and then when
the cache cleaner runs it will try to adjust the cache
size down to 0 again.

/robo

[Rabbit-proxy-users] How to compress but not cache?

[Costa T. <kt...@ho...>]

Hi,

I am going to have RabbIT2 between two other HTTP caches. How can I make
it compress but not cache anything at all?

When I use the DontCacheFilter it seems that it does not convert images
(although it compresses HTMLs).

Any help is appreciated.

Costa

PS: tested with v2.0.31

Re: [Rabbit-proxy-users] Weird Cache 0 byte problem again

[Sam <Sam...@Co...>]

Seems to have been a temporary problem.
I got a few reports and duplicated it myself but then it became "solved".

Don't forget, I am not using cache with Rabbit.
Still cannot for some reason but continue to work on the problem.

I still get high, too high, memory usage when using cache.
The bigger the cache the bigger the ram (1 GIG cache 256 megs ram, etc).

With not using cache and setting the cleanloop to 600 I can get a good 
24 hours before it reaches 512 megs ram.
I also tell Rabbit it can use 512 megs RAM. I could add more ram but 
that would be pointless because I would only gain more time.

I have a cron job that at 4:00 am it will "killall -9 java" and then 
restart the proxy.
Bad but better than doing it every hour since it takes a few minutes for 
the proxy to "go" on startup.

Sam

Robert Olofsson wrote:

> Samuel Hill wrote:
>
>> For example, when you go to yahoo and click on mail I see this in the
>> logs...
>>
>> 63.170.141.40 - - 07/May/2004:15:33:09 GMT "GET
>> http://www.yahoo.com/_ylh=X3oDMTB1M2EzYWFoBF9TAzI3MTYxNDkEdGVzdAMwBHRtcG
>> wDaWUtYmV0YQ--/r/m1 HTTP/1.1" 200 0
>> But the page never shows up in explorer.
>> Before I saw this with images but in this case it is the whole page.
>>  
>>
> Odd. yahoo mail is one of the things that I always test rabbit with.
> What I can see the yahoo.com pages are not cached (at least not when 
> mozilla is used).
> The images on yahoo are cached, but not the main page.
>
> I get something like:
> "GET 
> http://www.yahoo.com/_ylh=X3oDMTB1c2ZmZzF2BF9TAzI3MTYxNDkEdGVzdAMwBHRtcGwDbnMtYmV0YQ--/r/m1 
> HTTP/1.1" 200 -
>
> Can you check if yahoo.com is in rabbits cache? (any page in yahoo).
> Can you check if yahoo.com is in IE's cache?  (or remove temporary 
> internet files and try again).
>
> It may be that your browser tries a conditional request and does not 
> handle the
> response correctly.
>
> /robo
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by Sleepycat Software
> Learn developer strategies Cisco, Motorola, Ericsson & Lucent use to 
> deliver higher performing products faster, at low TCO.
> http://www.sleepycat.com/telcomwpreg.php?From=osdnemail3
> _______________________________________________
> Rabbit-proxy-users mailing list
> Rab...@li...
> https://lists.sourceforge.net/lists/listinfo/rabbit-proxy-users
>
>

Re: [Rabbit-proxy-users] Weird Cache 0 byte problem again

[Robert O. <ro...@kh...>]

Samuel Hill wrote:

>For example, when you go to yahoo and click on mail I see this in the
>logs...
>
>63.170.141.40 - - 07/May/2004:15:33:09 GMT "GET
>http://www.yahoo.com/_ylh=X3oDMTB1M2EzYWFoBF9TAzI3MTYxNDkEdGVzdAMwBHRtcG
>wDaWUtYmV0YQ--/r/m1 HTTP/1.1" 200 0 
>
>But the page never shows up in explorer.
>Before I saw this with images but in this case it is the whole page.
>  
>
Odd. yahoo mail is one of the things that I always test rabbit with.
What I can see the yahoo.com pages are not cached (at least not when 
mozilla is used).
The images on yahoo are cached, but not the main page.

I get something like:
 "GET 
http://www.yahoo.com/_ylh=X3oDMTB1c2ZmZzF2BF9TAzI3MTYxNDkEdGVzdAMwBHRtcGwDbnMtYmV0YQ--/r/m1 
HTTP/1.1" 200 -

Can you check if yahoo.com is in rabbits cache? (any page in yahoo).
Can you check if yahoo.com is in IE's cache?  (or remove temporary 
internet files and try again).

It may be that your browser tries a conditional request and does not 
handle the
response correctly.

/robo

[Rabbit-proxy-users] Weird Cache 0 byte problem again

[Samuel H. <Sam...@Co...>]

For example, when you go to yahoo and click on mail I see this in the
logs...

63.170.141.40 - - 07/May/2004:15:33:09 GMT "GET
http://www.yahoo.com/_ylh=X3oDMTB1M2EzYWFoBF9TAzI3MTYxNDkEdGVzdAMwBHRtcG
wDaWUtYmV0YQ--/r/m1 HTTP/1.1" 200 0 

But the page never shows up in explorer.
Before I saw this with images but in this case it is the whole page.

There is nothing in the error log.

Sam

Re[2]: [Rabbit-proxy-users] 504 Gateway Time-out

[Luis S. <lso...@gl...>]

That was it.  Setting ScrictHTTP=3Dfalse 
fixed the problem.

Thanks for your help.

--luis

*********** REPLY SEPARATOR  ***********

On 5/4/2004 at 8:10 AM Robert Olofsson wrote:

>Luis Soltero wrote:
>
>>any idea why this url fails?
>>http://www.ndbc.noaa.gov/station_page.phtml?station=3D42020
>>  
>>
>Do you run with strict http or not? "StrictHTTP=3Dtrue"
>I would guess that the web site sends broken http header that only works=
 if
>you do not use strict http.
>I have not checked yet, but I will, but the fact that it works for Sam
>seems to be an indication (he uses StrictHTTP=3Dfalse).
>
>Using strict http is in ways better and needed to pass the http test suite
>but many web sites sends broken http headers...
>
>The normal cause for this is that they something like a perl program that
>does this:
>
>print "Content-Type: text/html\n\n"
>print "<html......"
>
>This is broken the http line separator is \r\n and the above ought to 
>have two
>line separators after the content type.
>Both IE and mozilla happily accepts the broken format....
>
>/robo
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by: Oracle 10g
>Get certified on the hottest thing ever to hit the market... Oracle 10g. 
>Take an Oracle 10g class now, and we'll give you the exam FREE. 
>http://ads.osdn.com/?ad_id=3D3149&alloc_id=3D8166&op=3Dclick
>_______________________________________________
>Rabbit-proxy-users mailing list
>Rab...@li...
>https://lists.sourceforge.net/lists/listinfo/rabbit-proxy-users



Luis Soltero, Ph.D., MCS
Director of Software Development
Global Marine Networks, LLC
StarPlilot, LLC
Tel: 865-379-8723
Fax: 865-681-5017
E-Mail: lso...@gl...
Web: http://www.globalmarinenet.net
Web: http://www.starpilotllc.com

Wireless E-Mail, Web Hosting, Weather and more...
                                     and
StarPilot, the state of the art in navigation computations at your finger=
 tips...

Re: [Rabbit-proxy-users] 504 Gateway Time-out

[Robert O. <ro...@kh...>]

Luis Soltero wrote:

>any idea why this url fails?
>http://www.ndbc.noaa.gov/station_page.phtml?station=42020
>  
>
Do you run with strict http or not? "StrictHTTP=true"
I would guess that the web site sends broken http header that only works if
you do not use strict http.
I have not checked yet, but I will, but the fact that it works for Sam
seems to be an indication (he uses StrictHTTP=false).

Using strict http is in ways better and needed to pass the http test suite
but many web sites sends broken http headers...

The normal cause for this is that they something like a perl program that
does this:

print "Content-Type: text/html\n\n"
print "<html......"

This is broken the http line separator is \r\n and the above ought to 
have two
line separators after the content type.
Both IE and mozilla happily accepts the broken format....

/robo

Re: [Rabbit-proxy-users] 504 Gateway Time-out

[Sam <Sam...@Co...>]

What version of all are you using?
What version of Rabbit and what version of java.

It works fine in my setup but I have a pretty new version.

Sam

Luis Soltero wrote:

>any idea why this url fails?
>
>http://www.ndbc.noaa.gov/station_page.phtml?station=42020
>
>thanks,
>
>--luis
>
>
>Luis Soltero, Ph.D., MCS
>Director of Software Development
>Global Marine Networks, LLC
>StarPlilot, LLC
>Tel: 865-379-8723
>Fax: 865-681-5017
>E-Mail: lso...@gl...
>Web: http://www.globalmarinenet.net
>Web: http://www.starpilotllc.com
>
>Wireless E-Mail, Web Hosting, Weather and more...
>                                     and
>StarPilot, the state of the art in navigation computations at your finger tips...
>
>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by: Oracle 10g
>Get certified on the hottest thing ever to hit the market... Oracle 10g. 
>Take an Oracle 10g class now, and we'll give you the exam FREE. 
>http://ads.osdn.com/?ad_id149&alloc_id66&op=click
>_______________________________________________
>Rabbit-proxy-users mailing list
>Rab...@li...
>https://lists.sourceforge.net/lists/listinfo/rabbit-proxy-users
>
>
>  
>

[Rabbit-proxy-users] 504 Gateway Time-out

[Luis S. <lso...@gl...>]

any idea why this url fails?

http://www.ndbc.noaa.gov/station_page.phtml?station=3D42020

thanks,

--luis


Luis Soltero, Ph.D., MCS
Director of Software Development
Global Marine Networks, LLC
StarPlilot, LLC
Tel: 865-379-8723
Fax: 865-681-5017
E-Mail: lso...@gl...
Web: http://www.globalmarinenet.net
Web: http://www.starpilotllc.com

Wireless E-Mail, Web Hosting, Weather and more...
                                     and
StarPilot, the state of the art in navigation computations at your finger=
 tips...