>Therefore, I think JVM1 and JVM2 should be client node with near caches and also have node filter. Not really, you can choose only one solution here - NodeFilter or client nodes with Near cache. If you want to store data on nodes JVM1 and JVM2, then you should go with NodeFilter approach. In this case they will be server nodes and they will have all needed data locally. In case of using Near cache, entry will be added to it only after reading this entry from the remote cache first time. Also, this data will be removed after node restart.
Evgenii вт, 10 мар. 2020 г. в 06:20, Edward Chen <[email protected]>: > Here is a background. In enterprise wide, each application, like JVM1, > will create its own cache and this data is useful for another application, > so this data should push to server node , like JVM3. but each client node, > like JVM1, should have freedom to contain other caches based on some rules, > but definitely not every cache in JVM3. in this case, looks like Node > filter will work. > > Therefore, I think JVM1 and JVM2 should be client node with near caches > and also have node filter. > > My understanding is correct ? > > Thanks Stephen Darlington for the near cache idea. > > > On 3/9/2020 7:04 PM, Evgenii Zhuravlev wrote: > > Hi, > > You can use NodeFilter for caches. Please use this JavaDoc for > information: > https://www.javadoc.io/doc/org.apache.ignite/ignite-core/latest/org/apache/ignite/util/AttributeNodeFilter.html > > Example can be found here: > https://github.com/ezhuravl/ignite-code-examples/blob/master/src/main/java/examples/nodefilter/cache/CacheNodeFilterExample.java > > Evgenii > > пт, 6 мар. 2020 г. в 18:34, Edward Chen <[email protected]>: > >> Hello, >> >> I want to achieve this topology, do you know how to configure ? >> >> The critical parts are, cache2 in JVM2 should not be replicated or copied >> to JVM 1 . cache1 in JVM1 should not be replicated or copied to JVM 2 . >> JVM3 and JVM 4 are each other failed-over backup. >> >> Thanks. Ed >> >
