Received: (qmail 79928 invoked by uid 501); 22 Dec 2000 17:05:04 -0000
Message-Id: <20001222170504.79927.qmail@locus.apache.org>
Date: 22 Dec 2000 17:05:04 -0000
From: Phil Dietz <pedietz@west.com>
Reply-To: pedietz@west.com
To: submit@bugz.apache.org
Subject: IE does not cache images to the site.  Thats because ap_make_etag uses the inode for building the ETag.
X-Send-Pr-Version: 3.110

>Number:         7010
>Category:       protocol
>Synopsis:       IE does not cache images to the site.  Thats because ap_make_etag uses the inode for building the ETag.
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    apache
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   apache
>Arrival-Date:   Fri Dec 22 09:10:01 PST 2000
>Closed-Date:    Sat Jan 05 11:48:27 PST 2002
>Last-Modified:  Sat Jan 05 11:48:27 PST 2002
>Originator:     pedietz@west.com
>Release:        1.3.14
>Organization:
>Environment:
Cisco Local Director that sits in front of a bunch of mirrored web servers.
>Description:
IE does not cache images to the site.  Thats because ap_make_etag uses the inode for building the ETag.

The local director round robins the transactions.  When you click refresh on your browser, you will more than likely hit a different unit than before.  Since the file is on a different inode, its never cached.

Netscape does not use ETags so it doensnt appear to be affected.
>How-To-Repeat:
http://www.orderprocessing.net/webmon/imagetest/
>Fix:
The default_handler in http_core.c should make weaker Etags since it pushes static content.
Stronger Etags should be made for the other handlers such as mod_cgi, etc.
I think the problem is that Apache source has one ap_make_etag function for everything. After reading the w3.org doc, each handler in Apache should have it's own ap_make_etag function.
CGI's will make its Etag based of SQL fields, etc.
At the very least, a directive to turn off inodes would solve it.
Since all browsers and proxies look for Last-Modified, a directive to turn off Etags all-together, will also solve it.

Note: more and more web traffic directors are coming out besides Cisco's line.  Intel has a whole line of them to at: http://www.intel.com/network/solutions/e-commerce.htm
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open-closed
State-Changed-By: coar
State-Changed-When: Sat Jan  5 11:48:27 PST 2002
State-Changed-Why:
This has been addressed with the FileETag directive
added in version 1.3.23-dev.  It should appear in the
next release after 1.3.22.
>Unformatted:
 [In order for any reply to be added to the PR database, you need]
 [to include <apbugs@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!     ]