1 Reply Latest reply on Oct 2, 2014 3:42 AM by Radim Vansa

    SingleFileStore implementation and expected usage

    Andreas Öh Newbie

      Hello,

      I have a cache with two nodes and are using write through to a single file store. The entries in the cache are expected to last for about 14 days so the cache's file store are expected grow to some 200gb of data. If one node goes down I use state transfer between nodes but I'm experience very long startup times when a node is rebooted, mainly I think the disk io is the problem.

       

      Is the SingleFileStore expected to handle these kinds of file sizes? It seems to work fine for smaller sized file stores, like ~10Gb.

      I'm using Infinispan version: 6.0.1

       

      Thanks

      Andreas

        • 1. Re: SingleFileStore implementation and expected usage
          Radim Vansa Master

          SingleFileStore contains no rocket science, it's just a ConcurrentHashMap of keys -> entry value offsets. When preloading data, the file is sequentially read and this map is built naively.

           

          When you have cluster with multiple nodes and one of them goes down, you shouldn't do any preloading - the data should be really moved by state transfer. If you use local storage (as single file store, or LevelDB store), you should reboot the node with purge=true to remove old entries.

           

          I'd guess that it's the state transfer that takes a while - you can tune it with the chunkSize attribute and see if increasing that reduces the ST duration.