+1 on release.
I did load tests and checked the documentation looks good.
This candidate may run worse than recent hbases' under load. I got
NSREs and timeout failures when running with 8 concurrent clients using
PE but it completed PE random and sequential Writes with 4 concurrent
clients -- the test I use to figure whether to vote on a hadoop candidate.
This candidate would have been better with a fix for HBASE-505 included
but we can add that later and make a 0.1.1 in a week or so.
St.Ack
stack wrote:
Here is a third candidate (candidate count is zero-based). Candidate
1 bundle was corrupt.
Please download and try it out:
http://people.apache.org/~stack/hbase-0.1.0-candidate-2/
Does this candidate work for you? Do unit tests pass on your
hardware? Is the documentation sufficient?
Should we release this? I'll leave the vote close at Thursday, March
27th.
Thanks,
St.Ack
