[
https://issues.apache.org/jira/browse/PIG-1229?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ankur updated PIG-1229:
-----------------------
Attachment: jira-1229-final.test-fix.patch
Here is my understanding of what happens
1. The main thread in the JVM executing the test initializes MiniDFSCluster,
MiniMRCluster and HSQLDB server all in different threads.
2. The test setUp() method then executed to create table 'ttt' to which data
will be written by DBStorage() in the test.
3. Pig statements are then executed that spawn M/R job as a separate process
that tries to get a connection to the database and create a preparedStatement
for table 'ttt'. This fails sometimes as DB thread does NOT get a chance to
fully persist the table information and the exception is thrown from the
map-tasks as noted by Ashutosh.
The fix for this is to add a 5 sec sleep in setUp() method to give DB a chance
to persist table information. This alleviates the problem and test passes for
repeated multiple runs.
Note that Ideal fix would have been to do a busy wait for table creation
completion but i don't see a method in HSqlDB to do that.
> allow pig to write output into a JDBC db
> ----------------------------------------
>
> Key: PIG-1229
> URL: https://issues.apache.org/jira/browse/PIG-1229
> Project: Pig
> Issue Type: New Feature
> Components: impl
> Reporter: Ian Holsman
> Assignee: Ankur
> Priority: Minor
> Fix For: 0.8.0
>
> Attachments: jira-1229-final.patch, jira-1229-final.test-fix.patch,
> jira-1229-v2.patch, jira-1229-v3.patch, pig-1229.2.patch, pig-1229.patch
>
>
> UDF to store data into a DB
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
