Hi All,

I am using Derby as an embedded database within a Hadoop job to lookup IP geographic info.  
http://mpouttuclarke.wordpress.com/2010/12/10/java-embedded-db-for-ip2location-in-hadoop/

The problem is that Hadoop has an option called JVM sharing where more than one thread may be active in the JVM instance.  Since the embedded option only supports one thread at a time, I have had to turn off JVM sharing for my IP lookup job (by setting mapred.job.reuse.jvm.num.tasks to 1).

I have already tried embedding the network server and client with a localhost loopback and it made the code 10 times slower.  So if anyone has been able to overcome this limitation without using the network server and client I would be very interested.

Thanks,
Matt

iCrossing Privileged and Confidential Information
This email message is for the sole use of the intended recipient(s) and may contain confidential and privileged information of iCrossing. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.