locked
AppFabric Local Cache Misses RRS feed

  • Question

  • I have a web application which is very cache-heavy, I'm converting from System.Web.Caching to AppFabric.

    What determines how much memory is consumed by the local cache? Is it configurable? I have a very high objectCount and ttl for the local cache, but I still see quite a few sporadic local cache misses that go through to the distributed cache (evidenced from perfmon counters on the cache hosts). The misses are causing minor performance problems (up to 3 seconds) for offsite developers working over a VPN connecting to a distributed cache on our internal network.

    It's clear that the majority of cache requests are going to the local cache; only about a dozen or so out of several hundred cache requests miss the local cache. Application clients are running x64 and have a lot of available memory, yet the application experiences local cache misses for numerous cache requests. It's just using the standard Cache.Get(...), which I know hits the local cache.

    Any ideas? It appears like the local cache is experiencing evictions although there's memory to spare and ttls and objectCounts are very high.

    • Edited by JRS52 Friday, July 19, 2013 5:40 PM
    Friday, July 19, 2013 5:32 PM