hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Aleksandr Shulman <al...@cloudera.com>
Subject Re: [VOTE] The 1st HBase 0.98.0 release candidate is available for download
Date Mon, 03 Feb 2014 22:22:27 GMT
Sorry to join the party a little late. I have two pieces of feedback for
the RC. Because of 2.1 and 2.2, I am also -1 on this RC.

1. Smoke tests - I ran the standard internal smoke tests - they all passed.
2. Binary/API incompatibilities:
  1. Scan.setSmall(boolean) return type changed, which will cause a binary
incompatibility between 0.96 and 0.98
  2. HTableDescriptor.isDeferredLogFlush was removed, which will cause a
binary incompatibility between 0.96 and 0.98. We should add it back with
default response of 'true'
  3. KeyValue constructors removed (known issue, but we should make sure to
document)
  4. ServerName constructors removed (known issue, but we should make sure
to document)

For 2.1 and 2.2:


Exception from testScanMeta:  -> java.lang.NoSuchMethodError:
org.apache.hadoop.hbase.client.Scan.setSmall(Z)V


Exception from testIsDeferredLogFlush:  ->
java.lang.NoSuchMethodError:
org.apache.hadoop.hbase.HTableDescriptor.isDeferredLogFlush()Z


The full report of Public API diffs can be found here:
http://people.apache.org/~jmhsieh/hbase_jdiff_report-p-0.96-c-0.98/changes.html
I will continue examining the report and I'd encourage folks to do the same.



On Tue, Jan 28, 2014 at 9:59 AM, Andrew Purtell <apurtell@apache.org> wrote:

> On Mon, Jan 27, 2014 at 7:04 PM, tsuna <tsunanet@gmail.com> wrote:
>
> > Sorry I didn't mean to sink the RC just for this.
>
>
> Well I think we should not be obnoxious in that way even if unintended.
> However, HBASE-10433 would also have sunk the RC if it makes you feel
> better. :-)
>
>
>
> --
> Best regards,
>
>    - Andy
>
> Problems worthy of attack prove their worth by hitting back. - Piet Hein
> (via Tom White)
>



-- 
Best Regards,

Aleks Shulman
847.814.5804
Cloudera

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message