Thanks Vinod!
I went through all 31 fixes (Common: 8, HDFS: 6, MAPREDUCE: 8, YARN: 9) landed
on branch-2.6? now, and all outstanding patches (>=5KB) are listed below and
they are all bug fix.
Haven't found any high risky patches so far but I noticed we have 3 significant
fixes (MAPREDUCE-5948, MAPREDUCE-6481 and MAPREDUCE-6549) for LineRecordReader.
Among them, MAPREDUCE-5948 and MAPREDUCE-6481 are completed fixes and we will
ship them in 2.7.2, so I think there is no risk with them. MAPREDUCE-6549,
together with an open JIRA: MAPREDUCE-6558, sounds like fix the same issue for
uncompressed/compressed input file. I already ping the JIRA owners to get
suggestions on if we could release MAPREDUCE-6549 independently or together
with MAPREDUCE-6558 in next release.
I already pinged all owners for open JIRAs that targeted for 2.6.3 to move it
to 2.6.4. Based on the feedback from JIRAs (and voices here too), I will create
branch-2.6.3 in next 1 day or 2. Thanks!?
List of Outstanding Patches:
HADOOP-10134 [JDK8] Fix Javadoc errors caused by incorrect or illegal tags
in doc comments. (apurtell via stevel) 38K
HDFS-9289. Make DataStreamer#block thread safe and verify genStamp in
commitBlock. (Chang Li via zhz) 12K
HDFS-9470. Encryption zone on root not loaded from fsimage after NN
restart. (Xiao Chen via wang) 5K
MAPREDUCE-6273. HistoryFileManager should check whether summaryFile exists
to avoid FileNotFoundException causing HistoryFileInfo into MOVE_FAILED state.
(zhihai xu via devaraj) 5K
MAPREDUCE-5948. org.apache.hadoop.mapred.LineRecordReader does not handle
multibyte record delimiters well (Vinayakumar B, Rushabh Shah, and Akira
AJISAKA via jlowe) 20K
MAPREDUCE-6481. LineRecordReader may give incomplete record and wrong
position/key information for uncompressed input sometimes. (Zhihai Xu via
jlowe) 19K
MAPREDUCE-6549. multibyte delimiters with LineRecordReader cause duplicate
records (wilfreds via rkanter) 31K
YARN-2859. ApplicationHistoryServer binds to default port 8188 in
MiniYARNCluster.(Vinod Kumar Vavilapalli via xgong) 6K (with addendum)
YARN-4344. NMs reconnecting with changed capabilities can lead to wrong
cluster resource calculations (Varun Vasudev via jlowe) 12K
YARN-3925. ContainerLogsUtils#getContainerLogFile fails to read container
log files from full disks. (zhihai xu via jlowe) 7K
Thanks,
Junping
________________________________
From: Vinod Kumar Vavilapalli <vino...@apache.org>
Sent: Monday, November 30, 2015 9:23 PM
To: Junping Du
Cc: Sangjin Lee; mapreduce-...@hadoop.apache.org; yarn-...@hadoop.apache.org;
Hadoop Common; hdfs-...@hadoop.apache.org; Haohui Mai; Chris Trezzo
Subject: Re: continuing releases on Apache Hadoop 2.6.x
+1, this sounds good to me.
Thanks
+Vinod
On Nov 25, 2015, at 11:44 AM, Junping Du
<j...@hortonworks.com<mailto:j...@hortonworks.com>> wrote:
May be we should quickly go through landed patches in 2.6.3 list and postpone
any risky patches to 2.6.4?
