Re: HTTP Compression issues with IIS 5
From: Tony Cooke (TCooke_at_elders.com.au)
Date: 09/22/04
- Next message: Seth: "Re: ISP Redundancy"
- Previous message: David Boyer: "Re: How to give web directory write access to scripts but not everyone?"
- In reply to: WenJun Zhang[msft]: "RE: HTTP Compression issues with IIS 5"
- Next in thread: WenJun Zhang[msft]: "Re: HTTP Compression issues with IIS 5"
- Reply: WenJun Zhang[msft]: "Re: HTTP Compression issues with IIS 5"
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 22 Sep 2004 10:07:19 +0930
Hi WenJun.
My comments are below.
Thank you for replying to my thread.
All the best,
Tony
> Hi Tony,
Hi again.
> Please check out the following article. It's generally hitting the
> main cause of users experiencing cached content behavior when using
> HTTP compression.
>
> 319384 Pages Do Not Expire as Expected After Web Site Content Change
> http://support.microsoft.com/?id=319384
>
> To prevent dynamic content from being cached, we'd either disable
> HTTP Compression on dynamic pages - uncheck the 'Compress application
> files' checkbox or modify the HcCacheControlHeader metabase property.
>
> Follow the KB instructed, you can decrease HcCacheControlHeader to a
> small value or just set it to "max-age=0" to test. Except for the
> ADSUTIL script, using MetaEdit to open your metabase could be a more
> clear way to do this:
I have implemented the "fix" in http://support.microsoft.com/?id=319384 as
you requested and that has fixed the caching problem. I will leave
"max-age=0" as I don't believe that we will want the ASP page's HTML to be
cached client side. One question with this though, does this mean that any
related images that are embeded in the page are not cached as well?
I have installed MetaEdit 2.2 as well and that works very well. I wish I had
known about this before.
> After making the change of this property, you can run iisreset
> command to restart IIS services. Some additional points are:
>
> 1) In a site properties->HTTP Headers tab, enable Content Expiration
> and select 'Expire Immediately'. This will allow IIS set 'no-cache'
> in the cache control header of HTTP response and send it to client
> browsers.
Won't this expire all content on the client immediately including images,
javascript files, etc?
This could defeat the purpose of having compressed content if we have to
resend everything every time.
> 2) After making all these changes at server-side, a current cached
> file may still remain in client Internet Temp folder, means IE may
> keep using the local copy and doesn't send a new request to IIS until
> 24 hours later. Therefore it may be necessary to clear Temporary
> Internet files in IE->Internet Options.
OK.
> For the PDF being truncated behavior and problematic streamed doc,
> can you send some sample files to me? I'd like to reproduce the
> issues at my side and perform some deeper researching on it. I did
> see few cases of PDF file, sometimes the problem is on Adobe Reader's
> OLE control(hosted by IE as the actual web client) but not lies on
> IIS server. If you simply download the PDF file instead of opening
> them in browser(right-click to save as), will the download be
> successful? You can send files to my email address directly:
> v-wzhang@online.microsoft.com (Please remove online.)
OK. I will send the files directly to the above address (minus "online")
> For your last question, to a high volume web site, HTTP compression
> costs CPU and response time to save network bandwidth. So in case a
> site has enough bandwidth, enable compression doesn't make any sense.
> The following White paper contains some useful info to evaluate its
> effect:
> Web and Application Server Infrastructure - Performance and
> Scalability
> http://www.microsoft.com/technet/prodtechnol/windowsserver2003/technol
> ogies/webapp/iis/iis6perf.mspx
True but the true test of a product under real life conditions, especially
your own product, is to use it on your own site. Surely these problems
wouldn't exist if Microsoft used it at least on a sub-section of their site.
Also, the page loads would be considerably faster if it was implemented,
especially on some of Microsoft's large text based areas. How can Microsoft
improve on something and promote it if they don't use it themself?
> Hope some info above can help. Please feel free to let me know if you
> meet any problem.
> Best regards,
Thank you again. Your help in this matter is greatly appreciated.
> WenJun Zhang
> Microsoft Online Support
> This posting is provided "AS IS" with no warranties, and confers no
> rights.
> Get Secure! - www.microsoft.com/security
>
- Next message: Seth: "Re: ISP Redundancy"
- Previous message: David Boyer: "Re: How to give web directory write access to scripts but not everyone?"
- In reply to: WenJun Zhang[msft]: "RE: HTTP Compression issues with IIS 5"
- Next in thread: WenJun Zhang[msft]: "Re: HTTP Compression issues with IIS 5"
- Reply: WenJun Zhang[msft]: "Re: HTTP Compression issues with IIS 5"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|