Forum Discussion

NZ_David_20489's avatar
NZ_David_20489
Icon for Nimbostratus rankNimbostratus
Nov 20, 2012

F5 LTM caching

Hi,

 

I've used the same HTTP profile across several Virtual Servers (effectively the standard HTTP-wan-optimized-compression-caching). A couple of Virtual Servers use the same URIs, but obviously these are different URLs. They also use common members (same IP and Port).

 

We seem to be having issues with cached content, old content is being delivered to the client. I have tested by removing the HTTP profile, this resolves the issue.

 

I would like to understand how is the cache maintained? Is there a separate cache per Virtual Server, per http Profile, or is it just based upon the URI or URL?

 

Also is there any chance a cached entry could cross Virtual Servers in the scenario mentioned above?

 

 

 

config
design

4 Replies

  • hoolio's avatar
    hoolio
    Icon for Cirrostratus rankCirrostratus
    Nov 20, 2012
    Hi David,

     

     

    RAM cache is per profile not per VS. If you want a separate cache per VS, you'd want to create a separate profile.

     

     

    Aaron
  • uni's avatar
    uni
    Icon for Altocumulus rankAltocumulus
    Nov 22, 2012
    So, for any given profile, is the cache table indexed by the URI only, i.e. without the host?
  • Michael_Yates's avatar
    Michael_Yates
    Icon for Nimbostratus rankNimbostratus
    Nov 23, 2012
    The Host Value is included in a RAM Cache Entry.

     

     

    You can see all of the information that is included for each entry in the iControl API Reference for RAMCacheEntry