squid and byte ranges

We’re still having trouble with the “range_offset_limit” being set to -1.  Every day or so I get a long block of time where our bandwidth is entirely consumed by squid download a file that no client is reading.  I mentioned this directive in the last post.  What it does is force squid to download the entirety of a file any time a byte range of the file is requested.  When caching windows updates, this is very important to make sure the update actually gets cached.  It does, however, have some nasty side effects.

One of the biggest problems is that I can’t even filter it by firewall, apparently.  What I wanted to do was set up an iptables rule that said if a connection was being made where the destination was part of msecn.com or llnw.net, don’t run it through squid.  The problem is that some of the windows updates are actually coming from LimeLight (llnw.net).  I’m willing to bet that if I check, some are coming from msecn as well.

Right now most of our most used updates are already cached, so commented out my range_offset_limit setting.  Recently, though, we’ve been getting a few more vista machines.  These updates don’t seem to be getting cached right.  I have many requests for specific updates, and most of them are misses.  I just added .psf to the file filter on my refresh patterns yesterday, but I would think the two identical vista computers I downloaded updates for today would have gotten a few from the cache.

I guess I’ll just have to watch it and see.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: