production environment - Shared java HashMap in a clustered enviroment -

i've client application requesting information url every 1 second.

in server (a servlet & jsp application), in order avoid db access when it's not necessary, it's been implemented next solution. here's snippet:

//a static hashmap save last record inserted in db public static map<long, long> values = new hashmap<long, long>();   // lastrecordread sent client  if (values.get(id) != lastrecordread) {          //access database information      //cause last value read different last record inserted     ... }else{     //do nothing     //it's not necessary access db cause parameters match } 

this working expected in development environment.

the problem comes when have clustered environment. have server deployed in 2 nodes (using jboss), each 1 own hashmap, , own values. depending on node attack, can different values...

¿is there way share hashmap between 2 nodes? looking answer don't need keep 2 maps updated, meaning not calls between nodes...

any kind of appreciated.

edit: i'm playing hazelcast, , seems easy i'm afraid i'm doing wrong...

in server i'm using hazelcast instead of hasmap:

public static map<long, long> values = (hazelcast.newhazelcastinstance(new config())).getmap("values");  

when records inserted:

        if (((values.get(id) == null)||(values.get(id) < lastidinserted))) {             values.put(id, lastidinserted);              } 

when server called client app:

// lastrecordread sent client  if (values.get(id) != lastrecordread) {          //access database information      //cause last value read different last record inserted     ... }else{     //do nothing     //it's not necessary access db cause parameters match } 

and think, that's all. confirm if ok or missing something..? solution populating on nodes? i've been doing tests 2 tomcats , work, work differents ips?

thanks in advance!

you have 2 options:

1. use distributed key-value like: http://memcached.org/

http://infinispan.org/about/

http://basho.com/riak/

http://hazelcast.org/

...and tons of others.

2. use 'publisher-subscriber' concept , update each hashmap instance events. typically implemented jms broker:

   http://docs.oracle.com/cd/e19717-01/819-7759/aerbk/index.html https://www.rabbitmq.com/tutorials/tutorial-three-java.html

the choice depends on needs: fastest read , seek, without networking delays slow updates - use second option. solution data not changing frequently: geographical names, addresses , on.

as general case - use fist one.