Why the cached data lost from Coherence?

Question

Yesterday,I sarted a storage-enabledcache server by running cache-server.cmd and started a storage-disabled cache console by running coherence.cmd.
Then I inputed the following commands to input some data into the Coherence caches:
cache(cb)
put a a
put b b
put c c
put d d
put e e
Map (cb): list
c = c
e = e
a = a
d = d
b = bBut today, when I issue the list command from the Coherence command line, the output become "null".
Why the data lost?

Answer 1

By running cache-server.cmd Coherence will be using the configuration file from the coherence.jar which will default to a cache with expiry of one hour. So if you had checked you cache after an hour or so you would have seen it was empty.
If you look at the coherence-cache-config.xml file inside the coherence jar you will see this mapping (which will apply to the cache you created with the name "cb"
    <cache-mapping>
      <cache-name>*</cache-name>
      <scheme-name>example-distributed</scheme-name>
    </cache-mapping>This maps to a scheme called example-distributed that looks like this:
    <distributed-scheme>
      <scheme-name>example-distributed</scheme-name>
      <service-name>DistributedCache</service-name>
      <!-- To use POF serialization for this partitioned service,
           uncomment the following section -->
      <!--
      <serializer>
        <class-name>com.tangosol.io.pof.ConfigurablePofContext</class-name>
      </serializer>
      -->
      <backing-map-scheme>
        <local-scheme>
          <scheme-ref>example-binary-backing-map</scheme-ref>
        </local-scheme>
      </backing-map-scheme>
      <autostart>true</autostart>
    </distributed-scheme>which uses the example-binary-backing-map scheme
    <local-scheme>
      <scheme-name>example-binary-backing-map</scheme-name>
      <eviction-policy>HYBRID</eviction-policy>
      <high-units>{back-size-limit 0}</high-units>
      <unit-calculator>BINARY</unit-calculator>
      <expiry-delay>{back-expiry 1h}</expiry-delay>
      <flush-delay>1m</flush-delay>
      <cachestore-scheme></cachestore-scheme>
    </local-scheme>which you can see has expiry set to 1 hour.
I once had to help out a team who managed to get a node using this configuration on their production cluster and lost random data every hour!
JK