Hi,

I have noticed an unusual pattern on static file requests in our site
(IIS6/Win2k3). The server was sending not modified files to the client , even
if the if-modified-since header was correct. I found these links on the web
describing a bug in IIS. The Etag header was not handled correctly and the
304 Not Modified answerd was not sent. Instead, IIS sends the file (200 OK
code).

http://dotnetjunkies.com/WebLog/leon/archive/2005/02/16/54630.aspx
http://support.microsoft.com/default.aspx?scid=kb;EN-US;823544
http://support.microsoft.com/default.aspx?scid=kb;EN-US;900245

Is it safe to use a filter that removes the Etag from all requests to the
site as suggested in
http://blogs.msdn.com/yanno/archive/2006/03/03/543176.aspx ?

Thanks,

IIS Cache Etag bug by Sergio

Sergio
Sat Apr 29 08:27:01 CDT 2006

I undestand that the Etag header is used to improve cache control mechanism.
However it is not working, IIS6/Win2k3 keeps changing the Etag even to not
modified static files. The browser sends the correct if-modified-since date
(almost 2 years ago) and the Etag received from the server. IIS generates a
new Etag for this same file and replies with 200 OK instead of 304 Not
Modified header.

We have an intranet providing service to one client and a huge bandwidth
waste due to this bug, tons of GB/Month in static files. So, is it safe to
use a filter to remove the etag from the headers sent by IIS while we wait
for a fix from Microsoft?

"Sergio" wrote:

> Hi,
>
> I have noticed an unusual pattern on static file requests in our site
> (IIS6/Win2k3). The server was sending not modified files to the client , even
> if the if-modified-since header was correct. I found these links on the web
> describing a bug in IIS. The Etag header was not handled correctly and the
> 304 Not Modified answerd was not sent. Instead, IIS sends the file (200 OK
> code).
>
> http://dotnetjunkies.com/WebLog/leon/archive/2005/02/16/54630.aspx
> http://support.microsoft.com/default.aspx?scid=kb;EN-US;823544
> http://support.microsoft.com/default.aspx?scid=kb;EN-US;900245
>
> Is it safe to use a filter that removes the Etag from all requests to the
> site as suggested in
> http://blogs.msdn.com/yanno/archive/2006/03/03/543176.aspx ?
>
> Thanks,

Re: IIS Cache Etag bug by anilruia

anilruia
Sat Apr 29 11:12:50 CDT 2006

By default IIS updates the e-tag whenever there is a change in the
metabase (which can happen without any user invoked action), you can
use the MD_ETAG_CHANGE_NUMBER property to have the e-tag not change on
metabase changes - just set it to some fixed value -
http://technet2.microsoft.com/WindowsServer/en/Library/f838f2c2-ab20-441c-8893-570cfc40156f1033.mspx


Re: IIS Cache Etag bug by Sergio

Sergio
Sat Apr 29 13:38:01 CDT 2006

There is an article at MSDN that states: "When you set the
MD_ETAG_CHANGENUMBER metabase property in Microsoft Internet Information
Services version 6.0 to improve Web browser caching performance, performance
does not improve." The link to the article:
http://support.microsoft.com/default.aspx?scid=kb;EN-US;823544

Did you set a value to this property in your server ? Did it work for you ?
Are you using Win2k3 with Service Pack 1?

-------------------------------------------------------------------
"anilruia@gmail.com" wrote:

> By default IIS updates the e-tag whenever there is a change in the
> metabase (which can happen without any user invoked action), you can
> use the MD_ETAG_CHANGE_NUMBER property to have the e-tag not change on
> metabase changes - just set it to some fixed value -
> http://technet2.microsoft.com/WindowsServer/en/Library/f838f2c2-ab20-441c-8893-570cfc40156f1033.mspx
>
>