Squid Web Cache wiki

Squid Web Cache documentation

🔗 Caching Windows Updates

by Yuri Voinov

🔗 Outline

Windows Updates is one of most common task for caching.

🔗 Usage

The Alternative is using WU server. When it’s impossible for some reasons your can use Squid to cache this.

🔗 Squid Configuration File

Paste the configuration file like this:

# Updates: Windows
refresh_pattern -i windowsupdate.com/.*\.(cab|exe|ms[i|u|f|p]|[ap]sf|wm[v|a]|dat|zip|psf) 43200 80% 129600 reload-into-ims
refresh_pattern -i microsoft.com/.*\.(cab|exe|ms[i|u|f|p]|[ap]sf|wm[v|a]|dat|zip|psf) 43200 80% 129600 reload-into-ims
refresh_pattern -i windows.com/.*\.(cab|exe|ms[i|u|f|p]|[ap]sf|wm[v|a]|dat|zip|psf) 43200 80% 129600 reload-into-ims
refresh_pattern -i microsoft.com.akadns.net/.*\.(cab|exe|ms[i|u|f|p]|[ap]sf|wm[v|a]|dat|zip|psf) 43200 80% 129600 reload-into-ims
refresh_pattern -i deploy.akamaitechnologies.com/.*\.(cab|exe|ms[i|u|f|p]|[ap]sf|wm[v|a]|dat|zip|psf) 43200 80% 129600 reload-into-ims

🔗 Troubleshooting

🔗 How do I make Windows Updates cache?

Windows Update generally (but not always) uses HTTP Range-Offsets’ (AKA file partial ranges) to grab pieces of the Microsoft Update archive in parallel or using a random-access algorithm trying to reduce the web traffic. Some versions of Squid do not handle or store Ranges very well yet.

A mix of configuration options are required to force caching of range requests. Particularly when large objects are involved.

:warning: Windows 8.1 upgrade pack requires up to 5GB objects to be cached. It will however, cache nicely provided the size limit is set high enough.

:information_source: Due to the slow-down problem below we recommend service packs be handled specially: Extend the maximum cached object size to the required size, then run a full download on a single machine, then run on a second machine to verify the cache is being used. Only after this verification succeeds open updating to all other machines through the proxy.

🔗 Preventing Early or Frequent Replacement

Once you have done the above to cache updates you encounter the problem that some software often forces a full object reload instead of revalidation. Which pushes the cached content out and fetches new objects very frequently.

An idea that was floating around suggested that you use a refresh_pattern regexp config to do your WU caching. I decided to test this idea out in my squid proxy, along with one or 2 other ideas (the other ideas failed hopelessly but the WU caching worked like a charm.)

The idea basically suggested this:

refresh_pattern microsoft.com/.*\.(cab|exe|ms[i|u|f]|asf|wm[v|a]|dat|zip) 4320 80% 43200

The original idea seemed to work in theory, yet in practicality it was pretty useless - the updates expired after 30 minutes, there were download inconsistencies, and a whole array of issues. So looking at the HTTP responses and documentation for refresh_pattern, there was an extra clause that could be added. This is how it changed:

refresh_pattern -i microsoft.com/.*\.(cab|exe|ms[i|u|f]|[ap]sf|wm[v|a]|dat|zip) 4320 80% 43200 reload-into-ims
refresh_pattern -i windowsupdate.com/.*\.(cab|exe|ms[i|u|f]|[ap]sf|wm[v|a]|dat|zip) 4320 80% 43200 reload-into-ims

Now all that this line tells us to do is cache all .cab, .exe, .msu, .msu, .msf, .asf, .psf, .wma,….. to .zip from microsoft.com, and the lifetime of the object in the cache is 4320 minutes (aka 3 days) to 43200 minutes (aka 30 days). Each of the downloaded objects are added to the cache, and then whenever a request arrives indicating the cache copy must not be used it gets converted to an if-modified-since check instead of a new copy reload request.

So adding it to the original Squid settings to do with refresh_pattern, we get:

# Add one of these lines for each of the websites you want to cache.
refresh_pattern -i microsoft.com/.*\.(cab|exe|ms[i|u|f]|[ap]sf|wm[v|a]|dat|zip) 4320 80% 43200 reload-into-ims
refresh_pattern -i windowsupdate.com/.*\.(cab|exe|ms[i|u|f]|[ap]sf|wm[v|a]|dat|zip) 4320 80% 43200 reload-into-ims
refresh_pattern -i windows.com/.*\.(cab|exe|ms[i|u|f]|[ap]sf|wm[v|a]|dat|zip) 4320 80% 43200 reload-into-ims

refresh_pattern ^ftp:           1440    20%     10080
refresh_pattern ^gopher:        1440    0%      1440
refresh_pattern -i (/cgi-bin/|\?) 0     0%      0
refresh_pattern .               0       20%     4320

This should limit the system from downloading windows updates a trillion times a minute. It’ll hand out the Windows updates, and will keep them stored in the squid cache.

I also recommend a 30 to 60GB cache_dir size allocation, which will let you download tonnes of windows updates and other stuff and then you won’t really have any major issues with cache storage or cache allocation or any other issues to do with the cache.

🔗 Why does it go so slowly through Squid?

The work-around used by many cache maintainers has been to set the above config and force Squid to fetch the whole object when a range request goes through.

:information_source: Compounding the problem and ironically causing some slowdowns is the fact that some of the Microsoft servers may be telling your Squid not to store the archive file. This means that Squid will pull the entire archive every time it needs any small piece.

You will need to test your squid config with smaller values for the range_offset_limit bypass and see which provides the best results for you.

Another symptoms which occasionally appear when attempting to force caching of windows updates is service packs.

:information_source: If the quick_abort_min, quick_abort_max, quick_abort_pct settings are set to abort a download incomplete and a client closes with almost but not quite enough of the service pack downloaded. That clients following requests will often timeout waiting for Squid to re-download the whole object from the start. Which naturally causes the problem to repeat on following restart attempts.

🔗 How do I stop Squid popping up the Authentication box for Windows Update?

Add the following to your squid.conf, assuming you have defined localnet to mean your local clients. It ‘MUST’ be added near the top before any ACL that require authentication.

acl windowsupdate dstdomain windowsupdate.microsoft.com
acl windowsupdate dstdomain .update.microsoft.com
acl windowsupdate dstdomain download.windowsupdate.com
acl windowsupdate dstdomain redir.metaservices.microsoft.com
acl windowsupdate dstdomain images.metaservices.microsoft.com
acl windowsupdate dstdomain c.microsoft.com
acl windowsupdate dstdomain www.download.windowsupdate.com
acl windowsupdate dstdomain wustat.windows.com
acl windowsupdate dstdomain crl.microsoft.com
acl windowsupdate dstdomain sls.microsoft.com
acl windowsupdate dstdomain productactivation.one.microsoft.com
acl windowsupdate dstdomain ntservicepack.microsoft.com

acl wuCONNECT dstdomain www.update.microsoft.com
acl wuCONNECT dstdomain sls.microsoft.com

http_access allow CONNECT wuCONNECT localnet
http_access allow windowsupdate localnet

The above config is also useful for other automatic update sites such as Anti-Virus vendors, just add their domains to the acl.

:information_source: If you have squid listening on a localhost port with other software in front (ie dansGuardian). You will probably need to add permission for localhost address so the front-end service can relay the requests.

http_access allow CONNECT wuCONNECT localnet
http_access allow CONNECT wuCONNECT localhost
http_access allow windowsupdate localnet
http_access allow windowsupdate localhost

🔗 Squid problems with Windows Update v5

🔗 AKA, Why does Internet Explorer work but the background automatic updates fail?

There seems to be some problems with Microsoft Windows to access the Windows Update website. This is especially a problem when you block all traffic by a firewall and force your users to go through a proxy.

Symptom: Windows Update gives error codes like 0x80072EFD and cannot update, automatic updates aren’t working too.

Cause: In earlier Windows-versions Windows Update takes the proxy-settings from Internet Explorer. Since XP SP2 this is not sure. At my machine I ran Windows XP SP1 without Windows Update problems. When I upgraded to SP2 Windows Update started to give errors when searching updates etc.

The problem was that WU did not go through the proxy and tries to establish direct HTTP connections to Update-servers. Even when I set the proxy in IE again, it didn’t help . It isn’t Squid’s problem that Windows Update doesn’t work, but it is in Windows itself. The solution is to use the ‘proxycfg’ or ‘netsh’ tool shipped with Windows. With this tool you can set the proxy for WinHTTP.

:information_source: Similar issues are found with other Microsoft products in the same Windows versions. The commands below often fix all Microsoft proxy issues at once.

🔗 Proxy configuration with proxycfg

:information_source: In Windows Vista, Server 2008 and later proxycfg is obsolete. Use netsh instead.


C:\> proxycfg
# gives information about the current connection type. Note: 'Direct Connection' does not force WU to bypass proxy

C:\> proxycfg -d
# Set Direct Connection

C:\> proxycfg -p wu-proxy.lan:8080
# Set Proxy to use with Windows Update to wu-proxy.lan, port 8080

C:\> proxycfg -u
# Set proxy to Internet Explorer settings.

🔗 Proxy configuration with netsh


netsh winhttp set proxy ProxyName:80 "<local>"

C:\> netsh winhttp set proxy "localhost;"

To reset proxy settings for WinHTTP use:

C:\> netsh winhttp reset proxy

🔗 Squid with SSL-Bump and Windows Updates

In modern setups with Squid, Windows Update cannot be check updates with error “WindowsUpdate_80072F8F” or similar.

WU now uses its own pinned SSL certificate and must be spliced to work. When you use sniffer, you can see many IP’s with relatively big subnetworks. This leads to problems with a Squid-3.4 and causes serious problems when using Squid-3.5 or above.

To use splicing, you need to know the names of the servers, however, a recursive DNS query does not give a result.

To pass WU check through Squid splice, you only need to splice next MS servers:


For use in real setups, write file url.nobump:

# WU (Squid 3.5.x and above with SSL Bump)
# Only this sites must be spliced.

Just add this file as Squid ACL as follows:

acl DiscoverSNIHost at_step SslBump1
acl NoSSLIntercept ssl::server_name_regex -i "/usr/local/squid/etc/url.nobump"
ssl_bump splice NoSSLIntercept
ssl_bump peek DiscoverSNIHost
ssl_bump bump all

and you do not need to know all the IP authorization server for updates.

:information_source: NOTE: In some countries WU can product SQUID_X509_V_ERR_DOMAIN_MISMATCH error via Akamai. To do WU, you can require to add this into your Squid’s config:

    acl BrokenButTrustedServers dstdomain "/usr/local/squid/etc/dstdom.broken"
    acl DomainMismatch ssl_error SQUID_X509_V_ERR_DOMAIN_MISMATCH
    sslproxy_cert_error allow BrokenButTrustedServers DomainMismatch
    sslproxy_cert_error deny all

and add this to **dstdom.broken**:


:information_source: NOTE: Depending your Squid’s configuration, you may need to change your Squid’s cipher configuration to this one:

    sslproxy_cipher HIGH:MEDIUM:RC4:3DES:!aNULL:!eNULL:!LOW:!MD5:!EXP:!PSK:!SRP:!DSS

and add this one to your bumped port's configuration:


3DES and RC4 required to connect to WU and - **attention!** - Skype
assets site.

:warning: Some updates cannot be cached due to splice above. Beware!

:warning: Adding 3DES and, especially, RC4, produces potentially weak ciphers via client and WU/Skype and some other sites. Be careful!

Microsoft KB

🔗 An example of refresh_pattern that is being used at OpnSense


⚠️ Disclaimer: Any example presented here is provided "as-is" with no support
or guarantee of suitability. If you have any further questions about
these examples please email the squid-users mailing list.

To the FAQ Index

Categories: ConfigExample

Navigation: Site Search, Site Pages, Categories, 🔼 go up