[Hadoop-common-dev] [VOTE] Hadoop 0.17.2 release candidate 1

I should also mention that it felt too strange just using md5 for
corruption detection, since it has been broken. So I used:

% gpg2 --print-mds hadoop-0.17.2.tar.gz

which produces md5 and other checksums:

hadoop-0.17.2.tar.gz: MD5 = 48 19 EC 4A ED 32 8F 69 9E EA E2 9F 42
80 0F 03
hadoop-0.17.2.tar.gz: SHA1 = 83DC 753F 91F7 0BC9 D2E3 018F 58C1
48F6 BAB4
F2AC
hadoop-0.17.2.tar.gz: RMD160 = 84F1 C039 5E28 2419 6A9A 01A9 33D4
8E93 F6C7
CC88
hadoop-0.17.2.tar.gz: SHA224 = 2A314A09 154E14B2 2F4A93BA 5E490ACF
0077E182
4EAD89B3 73C16271
hadoop-0.17.2.tar.gz: SHA256 = 5F4EB1A7 49BFA392 437B2C9D 22E8CEB3
E50E5A05
7428EB99 4B14D20C 3DDB5ADD
hadoop-0.17.2.tar.gz: SHA384 = 221D2EA7 080AA207 2E31F662 B9948A6F
4A3419DE
56A1E1C7 A2707D8E 4616AFE8 EFE03537
DB350A81
D6551ECE 0637CD75
hadoop-0.17.2.tar.gz: SHA512 = 24F0DD8F 96C23145 07F25CA0 CB8853B1
61F21643
C0B7C299 84FF82FA 8F9B06CD D39F56DB
82A0E9CD
249C694E 034AD9D7 F5DD7EC9 EFB35302
86D49664
9D5124F4

and included it all in the md5 file. Does that break anything that
assumes the md5sum format? It is clearly very hard to break all 7 md's
at the same time. *smile* Should we do it for all of our releases?

-- Owen

+1

Have been using it today without issue doing hbase tests.

St.Ack


Owen O'Malley wrote:

+1 based on running some rudimentary tests and manifest checks. Note
that I have *not* run any benchmarks or other end-to-end tests on this
release.

Nige

+1. I have been running most of these patches (not all) for the last
month or so.

-dhruba

I've run the unit tests and a word count example on Mac OS.

+1

With 4 +1's and no -1's, the vote passes.

-- Owen