Hello All. I've setup an hbase (0.90.4) sandbox running on servers
where we have some excess capacity. Feel free to play with it, e.g.
create tables, run load tests, benchmarks, essentially do whatever you
want, just don't put your production services there, because while we
do have it up due to excess capacity, we may have to reclaim the
hardware at some point.
(don't worry about slamming it hard, those servers are running on
non-production zone of our network).
To create a client that can interface with the cluster, just download
hbase 0.90.4, and compile your code against the jars the come with it
(applies to the client below on paste bin).
I've created a very simple benchmark (proof of concept client) and
installed it in micro EC2 instance (you can find the code here
[ec2-user@ip-10-160-135-246 ~]$ java Hello 2>> /dev/null
Enter the number of rows you want to Put and Get : 10000
Enter the row value payload : 12345
Writing 10000 rows took 41470 milliseconds (<--- 4.1 ms per row not too bad!)
Reading 10000 rows took 43731 milliseconds
[ec2-user@ip-10-160-135-246 ~]$ hbase/bin/hbase shell
HBase Shell; enter 'help<RETURN>' for list of supported commands.
Type "exit<RETURN>" to leave the HBase Shell
Version 0.90.4, r1150278, Sun Jul 24 15:53:29 PDT 2011
1 row(s) in 0.9190 seconds
The Hbase zookeeper quorum has only one address
"img700.imageshack.us:2181", see code above on how to interface with
If you find this setup interesting or useful, or have questions about
please email me; Otherwise have fun!
(PS. Don't delete other people's tables and don't expose data you
don't want to be exposed, the cluster is read/write enabled for _all_,
we will carry no liabilities for anything whatsoever :)