hadoop-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Robert Evans <ev...@yahoo-inc.com>
Subject Re: [VOTE] 0.23.4 release
Date Mon, 01 Oct 2012 13:48:10 GMT
I think I would prefer to merge them onto branch-0.23 and have them come
out as part of 0.23.5.  It will probably show up in a month or so, as we
expect to find more issues as Yahoo! deploys this to more and larger
clusters.  If you really feel strongly about it I am OK with merging it
in, I just don't see a critical need for it right now.


On 9/30/12 4:28 PM, "Arun C Murthy" <acm@hortonworks.com> wrote:

>Hey Bobby,
> Thanks for getting this RC out.
> You might want to pull in YARN-137 and YARN-138 into this too.
>On Sep 29, 2012, at 8:35 AM, Robert Evans wrote:
>> I have built a 0.23.4-rc0 and am calling a release vote on it. As per
>>the bylaws this vote will be open for 1 week, and will close next
>>Saturday October 6th.
>> You can access the release through the apache maven staging repo, with
>>the 0.23.4-rc0 tag in subversion, and at
>> This release is a sustaining release with several important bug fixes
>>in it.  Most notably upgrading of a 1.0 cluster with append/sync enabled
>>should now work (HDFS-3731).  There are no known blockers against this
>>release, there are only two critical issues.
>> HADOOP-8811
>> Compile hadoop native library in FreeBSD
>> HDFS-3990
>> NN's health report has severe performance problems
>> HDFS-3990 only really has a significant impact on large clusters, and
>>can be mitigated by enabling DNS caching.
>> I have run several manual tests on this release on a small insecure
>>cluster and all of them pass I am a +1 (binding) for this release.
>> On a side note I am happy to announce that the 0.23 line is being
>>rolled out more widely at Yahoo!.  We just installed 0.23.3 on a 2000
>>node cluster and have only seen a few issues (HDFS-3990 is the biggest
>>of them).   I expect to do a 0.23.5 release in the coming weeks/months
>>as we find more issues running on large clusters.
>> --Bobby Evans
>Arun C. Murthy
>Hortonworks Inc.

View raw message