[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15473036#comment-15473036
]
Daniel Dai commented on PIG-4967:
-
That's fine for me. Patch committed to branch-0.16 as well.
> NPE in
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15464177#comment-15464177
]
Xiang Li commented on PIG-4967:
---
[~daijy] I set the fix version to 0.16.0, will you consider to commit this
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15454938#comment-15454938
]
Xiang Li commented on PIG-4967:
---
Hi Daniel, thanks very much for the review and guide! My pleasure to be a
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15448539#comment-15448539
]
Xiang Li commented on PIG-4967:
---
Daniel, thanks for the comments! I uploaded patch 3 with the following
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15448412#comment-15448412
]
Daniel Dai commented on PIG-4967:
-
I am fine with the fix. The line "Because MAPREDUCE-6762 fixes PIG-4967
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15447855#comment-15447855
]
Xiang Li commented on PIG-4967:
---
Hi, [~daijy], any comments?
> NPE in PigJobControl.run() when job status is
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15438597#comment-15438597
]
Xiang Li commented on PIG-4967:
---
Daniel, thanks for clarification! Got your idea.
I found it breaks UT (and
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15437217#comment-15437217
]
Daniel Dai commented on PIG-4967:
-
Sorry for confusion. When I say clone ControlledJob#toString(), I don't
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15436747#comment-15436747
]
Xiang Li commented on PIG-4967:
---
[~daijy] I reviewed all your previous comments and need to check with you if
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15436729#comment-15436729
]
Xiang Li commented on PIG-4967:
---
Looking into the exception which breaks UT against patch 1. It seems that I
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15436580#comment-15436580
]
Xiang Li commented on PIG-4967:
---
HI Daniel,
do you mean "ControlledJob.toString" rather than "Job.toString" in
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15436519#comment-15436519
]
Xiang Li commented on PIG-4967:
---
MAPREDUCE-6762 fixes this issue on Hadoop side. When status=null, the logic
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15433214#comment-15433214
]
Daniel Dai commented on PIG-4967:
-
I assume by doing " j.getJob().getStatus()==null", we don't need to pull
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15433011#comment-15433011
]
Xiang Li commented on PIG-4967:
---
Hi Daniel, sorry, patch 1 breaks some core set UT (launched by "ant
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15432529#comment-15432529
]
Xiang Li commented on PIG-4967:
---
[~daijy] Thanks for the comments.
Regarding
bq.Can we try to detect if
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15431866#comment-15431866
]
Daniel Dai commented on PIG-4967:
-
Thank for your patch, but it is longer than I expected. Can we try to
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15427814#comment-15427814
]
Xiang Li commented on PIG-4967:
---
[~daijy]
I uploaded the very first patch
1. It is almost to clone
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15424807#comment-15424807
]
Xiang Li commented on PIG-4967:
---
Hi Daniel, thanks for the explanation!
Regarding
bq. I am not sure what's
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15412955#comment-15412955
]
Daniel Dai commented on PIG-4967:
-
Decompose means instead of invoking j.toString, you can construct the
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15412884#comment-15412884
]
Xiang Li commented on PIG-4967:
---
Hi Daniel, thanks for the comments!
1. Could you please elaborate more on
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15412351#comment-15412351
]
Daniel Dai commented on PIG-4967:
-
Sure, you can decompose ControlledJob and do null check on individual
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15411975#comment-15411975
]
Xiang Li commented on PIG-4967:
---
Hadoop should fix this, but I think we can do something on Pig side.
> NPE
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15411967#comment-15411967
]
Xiang Li commented on PIG-4967:
---
We are using Pig 0.15.0 and Hadoop 2.7.1
> NPE in PigJobControl.run() when
[
https://issues.apache.org/jira/browse/PIG-4967?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15411966#comment-15411966
]
Xiang Li commented on PIG-4967:
---
Line 181 and 182 of PigJobControl are
{code}
ControlledJob j = it.next();
24 matches
Mail list logo