Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Zookeeper >> mail # user >> Using zookeeper server as library


Copy link to this message
-
Re: Using zookeeper server as library
Thanks for the info Rakesh.

On Tue, Mar 27, 2012 at 10:04 PM, Rakesh R <[EMAIL PROTECTED]> wrote:

> Hi Prakash,
>
> Following are the classes which has the logic of in-memory and persisting
> txn logs. It has the apis which would be doing respective functionalities.
> Please look into these, probably will get more ideas.
> In-mem structure : DataTree.java
> Persistence logic  : ZkDatabase.java, also refer
> org.apache.zookeeper.server.persistence for more...
>
> ZooKeeper is using the data elements as records (Record.java interface)
> and passing it to diff modules as entities.
> In your case, it would be CreateTxn, DeleteTxn, DataNode records etc.
>
> Refer this for details about jmx :
> http://zookeeper.apache.org/doc/r3.4.2/zookeeperJMX.html
>
> Best Regards,
> Rakesh R
> ________________________________________
> From: Prakash [[EMAIL PROTECTED]]
> Sent: Wednesday, March 28, 2012 12:26 AM
> To: [EMAIL PROTECTED]
> Subject: Using zookeeper server as library
>
> Hi all,
>
> I have a requirement, where my servers have to store a data structure
> similar to the zookepeer trees. Instead of writing it from scratch, i was
> wondering if i can just leverage the existing zookeeper code.
>
> The idea is to just link to the zookeeper.jar and use it locally via
> function calls. I dont need the client / server communication part. I dont
> need the replication part. I just need the code to create / delete nodes
> etc in memory. Persisting it in the log is a plus if can just leverage that
> too. The JMX stuff would be a bonus.
>
> Before i go try to do this, i wanted to know if you guys had any thoughts
> on this or if there any gotchas i need to look for.
>
> Thanks,
> -Prakash
>