[
https://issues.apache.org/jira/browse/PIG-2442?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13183422#comment-13183422
]
xuting zhao commented on PIG-2442:
----------------------------------
A few observations:
(1)If the multiple stores store independent variables, then it works fine.
For example,
DEFINE SCRIPT `a.pl` SHIP ('/homes/a.pl');
A = LOAD 'test.txt' USING PigStorage() ;
B1 = STREAM A THROUGH SCRIPT;
B2 = STREAM A THROUGH SCRIPT;
STORE B1 INTO 'B2' USING PigStorage();
STORE B2 INTO 'B2' USING PigStorage();
(2) The problem seems to be the ExecutableManager and the POstream which
only read one tuple into the I/O thread in the second stream operation and then
close.
I will look further into it. Anyone can give me some suggestion? :)
Xuting
> Multiple Stores in pig streaming causes infinite waiting
> --------------------------------------------------------
>
> Key: PIG-2442
> URL: https://issues.apache.org/jira/browse/PIG-2442
> Project: Pig
> Issue Type: Bug
> Affects Versions: 0.8.1, 0.9.0
> Reporter: Anitha Raju
> Fix For: 0.10
>
>
> Hi,
> If there are multiple store in a pig streaming script, it goes into infinite
> waiting.
> Script
> {code}
> DEFINE SCRIPT `./a.pl` SHIP ('/homes/anithar/a.pl');;
> DEFINE SCRIPT1 `./b.pl` SHIP ('/homes/anithar/b.pl');;
> A = LOAD 'test.txt' USING PigStorage() ;
> B1 = STREAM A THROUGH SCRIPT ;
> B1 = foreach B1 generate $0;
> STORE B1 INTO 'B1' USING PigStorage();
> B2 = STREAM B1 THROUGH SCRIPT1;
> STORE B2 INTO 'B2' USING PigStorage();
> {code}
> a.pl
> --------
> #! /usr/bin/perl -w
> while (my $line = <STDIN>) {
> print uc($line);
> }
> --------
> b.pl
> ---------
> #! /usr/bin/perl -w
> while (my $line = <STDIN>) {
> print $line;
> }
> ---------
> Input (test.txt)
> {code}
> test
> hi
> hello
> {code}
> This infinite waiting happens randomly causing the job to fail with "Task
> attempt failed to report
> status for 605 seconds. Killing!".
> Same happens with 0.8 version too.
> Regards,
> Anitha
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira