mod_proxy/10183: mod proxy failure with Pragma: no-cache URLs

[Andreas Steinmetz <as...@domdv.de>]

>Number:         10183
>Category:       mod_proxy
>Synopsis:       mod proxy failure with Pragma: no-cache URLs
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    apache
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Thu Mar 14 12:00:00 PST 2002
>Closed-Date:
>Last-Modified:
>Originator:     ast@domdv.de
>Release:        1.3.23
>Organization:
apache
>Environment:
Linux zeus 2.4.16 #2 Wed Mar 13 10:32:06 CET 2002 i686 unknown
>Description:
mod_proxy (with proxy and cache enabled) does cache URLs with "Pragma: no-cache" headers, i.e. the header with a content length of 0 bytes and thus no URL data is cached. On subsequent client requests this cache recors is found and an empty document (i.e. content length 0 bytes) is returned to the client.
This bug makes mod_proxy unusable for every page containing a "Pragma: no-cache" header.
>How-To-Repeat:
Below is a cut and paste example for http://www.kkcomputer.de/

Situation on zeus before first request and after first request:
===============================================================

zeus:/var/apache/proxy # ls -a
.  ..
zeus:/var/apache/proxy # ls -a
.  ..  .time  j
zeus:/var/apache/proxy #

Details of cache on zeus after first request:
=============================================

zeus:/var/apache/proxy # find . -ls
1155073    2 drwxr-xr-x   3 nobody   nogroup      2048 Mar 14 19:26 .
1155075    0 -rw-r--r--   1 nobody   nogroup         0 Mar 14 19:26 ./.time
7905281    1 drwx------   3 nobody   nogroup      1024 Mar 14 19:26 ./j
7913473    1 drwx------   3 nobody   nogroup      1024 Mar 14 19:26 ./j/P
7921665    1 drwx------   2 nobody   nogroup      1024 Mar 14 19:26 ./j/P/n
1155076    1 -rw-------   1 nobody   nogroup       617 Mar 14 19:26 ./j/P/n/qirWZTGaNwu5YpUFl@Q
zeus:/var/apache/proxy # cat ./j/P/n/qirWZTGaNwu5YpUFl@Q
000000003C90EB65 000000003C8A0703 000000003C919BBF 0000000000000002 000000003C90EB4F 000000003C90EB4F 0000000000000000
X-URL: http://www.kkcomputer.de/
Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, */*
Host: www.kkcomputer.de
Pragma: no-cache
User-Agent: curl/7.9.5 (i686-pc-linux-gnu) libcurl 7.9.5 (OpenSSL 0.9.6b)

HTTP/1.1 200 OK
Server: Microsoft-IIS/5.0
Content-Location: http://www.kkcomputer.de/index.html
Date: Thu, 14 Mar 2002 18:26:45 GMT
Content-Type: text/html
Accept-Ranges: bytes
Last-Modified: Sat, 09 Mar 2002 12:58:43 GMT
ETag: "80c425246ac7c11:a49"
Content-Length: 0

zeus:/var/apache/proxy #

Client situation:
=================

pcast2:/usr/src # curl -i -x zeus:8080 http://www.kkcomputer.de/
HTTP/1.1 200 OK
Date: Thu, 14 Mar 2002 18:26:11 GMT
Server: Apache
Accept-Ranges: bytes
Content-Length: 15653
Content-Location: http://www.kkcomputer.de/index.html
Content-Type: text/html
ETag: "80c425246ac7c11:a49"
Last-Modified: Sat, 09 Mar 2002 12:58:43 GMT
X-Cache: MISS from host001-server-1.lan.domdv.de

<html>

<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<title>K+K Computer GmbH N�rnberg</title>

[snip]

</body>

</html>pcast2:/usr/src # curl -i -x zeus:8080 http://www.kkcomputer.de/
HTTP/1.1 200 OK
Date: Thu, 14 Mar 2002 18:26:23 GMT
Server: Apache
Accept-Ranges: bytes
Content-Length: 0
Content-Location: http://www.kkcomputer.de/index.html
Content-Type: text/html
ETag: "80c425246ac7c11:a49"
Last-Modified: Sat, 09 Mar 2002 12:58:43 GMT
X-Cache: HIT from host001-server-1.lan.domdv.de (with revalidation)

pcast2:/usr/src #
>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
 [In order for any reply to be added to the PR database, you need]
 [to include <ap...@Apache.Org> in the Cc line and make sure the]
 [subject line starts with the report component and number, with ]
 [or without any 'Re:' prefixes (such as "general/1098:" or      ]
 ["Re: general/1098:").  If the subject doesn't match this       ]
 [pattern, your message will be misfiled and ignored.  The       ]
 ["apbugs" address is not added to the Cc line of messages from  ]
 [the database automatically because of the potential for mail   ]
 [loops.  If you do not include this Cc, your reply may be ig-   ]
 [nored unless you are responding to an explicit request from a  ]
 [developer.  Reply only with text; DO NOT SEND ATTACHMENTS!     ]