[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14059605#comment-14059605 ] Jonathan Eagles commented on PIG-3892: -- [~daijy], after this checkin for bin/pig script, now HADOOP_VERSION is always 1 since HADOOP_CORE_JAR is never zero length string. Perhaps a file exists check is in order. What do you think? {code} 296 HADOOP_CORE_JAR=`echo ${HADOOP_HOME}/hadoop-core*.jar` 297 298 if [ -z $HADOOP_CORE_JAR ]; then 299 HADOOP_VERSION=2 300 else 301 HADOOP_VERSION=1 302 fi {code} Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0, 0.14.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch, PIG-3892-3.patch, PIG-3892-4.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14059607#comment-14059607 ] Jonathan Eagles commented on PIG-3892: -- So sorry, I have jumped to a conclusion that was incorrect. This change is working for me. Apologies. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0, 0.14.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch, PIG-3892-3.patch, PIG-3892-4.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14038186#comment-14038186 ] Giridharan Kesavan commented on PIG-3892: - [~daijy] Thanks for not renaming the currently in use build targets. My only concern is about renaming the jar's {code} pig-0.13.1-SNAPSHOT-h2.jar - ./pig-0.13.1-SNAPSHOT.jar pig-0.13.1-SNAPSHOT-withouthadoop-h2.jar - pig-0.13.1-SNAPSHOT-withouthadoop.jar {code} maybe we can track that in a separate jira and see how we can deal with that. and Here is my +1 Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.14.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch, PIG-3892-3.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14014725#comment-14014725 ] Daniel Dai commented on PIG-3892: - Unlink it from 0.13.0 to unblock 0.13.0 release. We will still try get it into 0.13.0 if possible. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.14.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14012993#comment-14012993 ] Giridharan Kesavan commented on PIG-3892: - [~daijy] I will look into this patch later today. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14010828#comment-14010828 ] Daniel Dai commented on PIG-3892: - [~cos], jar/jar-withouthadoop/pigunit-jar/smoketests-jar will remain the same. I only change target package (and tar indirectly) Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14011230#comment-14011230 ] Cheolsoo Park commented on PIG-3892: {quote} How do you install Hadoop 2? I download Apache tar ball and it does not contain hadoop-core.jar. {quote} [~daijy], hadoop-core.jar does not exist in hadoop2. But the condition {{-z $HADOOP_CORE_JAR}} doesn't hold true when it doesn't exist since {{$HADOOP_CORE_JAR}} is always not be zero length. I think you should check whether hadoop-core.jar exists or not instead. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14011434#comment-14011434 ] Daniel Dai commented on PIG-3892: - Oh, I see what you mean. But we first set shopt -s nullglob in bin/pig, so the codesnip you posted is supposed to check existence of hadoop-core*.jar, right? I actually copied the code how we check existence of other jars (such as JYTHON_JAR at line 220), and tested works. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14011499#comment-14011499 ] Cheolsoo Park commented on PIG-3892: Thank you for the clarification. I didn't know about {{shopt -s nullglob}}. +1. Disclaimer: I assumes the Windows part of changes works. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14011794#comment-14011794 ] Konstantin Boudnik commented on PIG-3892: - bq. But we first set shopt -s nullglob in bin/pig that'd work of course, but I agree with other comment: doing {{[ -e -s ]}} might be less confusing. Not a show stopper though ;) Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14011836#comment-14011836 ] Daniel Dai commented on PIG-3892: - Ok, since there are many places in bin/pig use this patten, I will open a new ticket to do the refactory. Thanks for point out though. I would also like to wait for consent from [~gkesavan] before commit. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14010805#comment-14010805 ] Daniel Dai commented on PIG-3892: - bq. Whether hadoop1 or hadoop2, HADOOP_CORE_JAR is not zero-length, so the following doesn't work How do you install Hadoop 2? I download Apache tar ball and it does not contain hadoop-core.jar. Is there an easy way to distinguish h2 from h1 from your installation? bq. jar-all-h12 shouldn't depend on jar and jar-withouthadoop It should not, new patch will remove this bq.Won't this break BigTop for the same reason as PIG-3905 because ivy will resolve both hadoop1 and hadoop2 dependencies no matter what the value of hadoopversion is? Yes, ant package will create the distribution layout containing both pig-h1.jar and pig-h2.jar, which will need both hadoop 1 and hadoop 2 dependency. It might create some trouble if the private maven repository only carry h2 build artifacts. But can BIGTOP specify public maven repository as an alternative? [~gkesavan], [~cos], want to comment on this? Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14010815#comment-14010815 ] Konstantin Boudnik commented on PIG-3892: - During the build we sed'ing the {{ivy/ivysettings.xml}} to pick up artifacts from {{~/.m2/repository}} which are seeded from the Hadoop part of Bigtop build. We only build with {{-Dhadoopversion=23}} with explicit set of the targets like {{ jar jar-withouthadoop pigunit-jar smoketests-jar}} Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch, PIG-3892-2.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14008513#comment-14008513 ] Cheolsoo Park commented on PIG-3892: # Whether hadoop1 or hadoop2, {{HADOOP_CORE_JAR}} is not zero-length, so the following doesn't work- {code} +HADOOP_CORE_JAR=`echo ${HADOOP_HOME}/hadoop-core*.jar` + +if [ -z $HADOOP_CORE_JAR ]; then +HADOOP_VERSION=2 +else +HADOOP_VERSION=1 +fi {code} {code:title=hadoop1} HADOOP_CORE_JAR=/Users/cheolsoop/workspace/hadoop-1.0.3/hadoop-core-1.0.3.jar {code} {code:title=hadoop2} HADOOP_CORE_JAR='/Users/cheolsoop/workspace/hadoop-2.3.0/hadoop-core*.jar' {code} # jar-all-h12 shouldn't depend on {{jar}} and {{jar-withouthadoop}}- {code} +target name=jar-all-h12 depends=jar,jar-withouthadoop {code} # Won't this break BigTop for the same reason as PIG-3905 because ivy will resolve both hadoop1 and hadoop2 dependencies no matter what the value of {{hadoopversion}} is? {code} -target name=package depends=docs, api-report, jar, piggybank description=Create a Pig tar release +target name=package depends=jar-all-h12, docs, api-report, piggybank description=Create a Pig tar release {code} Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Attachments: PIG-3892-1.patch Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13971834#comment-13971834 ] Prashant Kommireddi commented on PIG-3892: -- Great, thanks. No use case, was wondering what the approach might be. Sounds good. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13969544#comment-13969544 ] Rohini Palaniswamy commented on PIG-3892: - We should go with 1. It is easy to do and one installation works for both Hadoop 1 and 2 depending upon what HADOOP_HOME or HADOOP_PREFIX points to. We just check for presence of hadoop-core*.jar in hadoop classpath and if present put pig-h1.jar in classpath else put pig-h2.jar in classpath. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13970001#comment-13970001 ] Alan Gates commented on PIG-3892: - +1 for 1. IIRC bin/hadoop has a -version option, so we don't even need to depend on magic jars being present, we can just ask hadoop. Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13970018#comment-13970018 ] Prashant Kommireddi commented on PIG-3892: -- +1 for 1. [~daijy] - would the way to invoke a certain version be passed as an argument to bin/pig, an env variable, both, something else? Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)
[jira] [Commented] (PIG-3892) Pig distribution for hadoop 2
[ https://issues.apache.org/jira/browse/PIG-3892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13970098#comment-13970098 ] Daniel Dai commented on PIG-3892: - It should be automatic. Do you have a use case user need to explicit pass the version number? Pig distribution for hadoop 2 - Key: PIG-3892 URL: https://issues.apache.org/jira/browse/PIG-3892 Project: Pig Issue Type: Bug Components: build Reporter: Daniel Dai Assignee: Daniel Dai Fix For: 0.13.0 Currently Pig distribution only bundle pig.jar for Hadoop 1. For Hadoop 2 users they need to compile again using -Dhadoopversion=23 flag. That is a quite confusing process. We need to make Pig work with Hadoop 2 out of box. I am thinking two approaches: 1. Bundle both pig-h1.jar and pig-h2.jar in distribution, and bin/pig will chose the right pig.jar to run 2. Make two Pig distributions for Hadoop 1 and Hadoop Any opinion? -- This message was sent by Atlassian JIRA (v6.2#6252)