[
https://issues.apache.org/jira/browse/PHOENIX-2074?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14613275#comment-14613275
]
Josh Mahonin commented on PHOENIX-2074:
---------------------------------------
I'm not sure if this is the same issue or not, but I'm running into something
similar with 4.4.0 with one of my spark jobs. This appears to be a regression
from 4.3.1. I don't yet have a lot of detail yet, but I am seeing the following
output from Spark:
java.lang.StackOverflowError
at
org.apache.phoenix.iterate.RoundRobinResultIterator$RoundRobinIterator.next(RoundRobinResultIterator.java:309)
at
org.apache.phoenix.iterate.RoundRobinResultIterator$RoundRobinIterator.next(RoundRobinResultIterator.java:309)
at
org.apache.phoenix.iterate.RoundRobinResultIterator$RoundRobinIterator.next(RoundRobinResultIterator.java:309)
It's a bit difficult to update the cluster I'm seeing this on right now, so I
won't be able to test out this patch readily until I can get another
environment up and load the dataset in. I can say that there are no JOINs
involved, as this is going through the spark -> mapreduce pipeline. There's a
few SELECTs happening at the same time, but my guess is it's failing on the one
dataset which is quite a lot larger than the others at 6632640 rows.
The SELECT itself isn't complicated, I believe it's getting translated to
something like:
SELECT ETYPE,EID,M,E FROM WD WHERE TID = '???';
Where the DDL is approximately:
CREATE TABLE IF NOT EXISTS WD(
TID CHAR(3) NOT NULL,
ETYPE CHAR(3) NOT NULL,
EID VARCHAR NOT NULL,
M INTEGER NOT NULL,
E DOUBLE
CONSTRAINT pk PRIMARY KEY (TID, ETYPE, EID, M)) MULTI_TENANT=true,
COMPRESSION='SNAPPY', KEEP_DELETED_CELLS = false;
I don't see the same issues with the same job, on a similar dataset where the
231840 rows. Note that it's hitting the same table, just with a different TID.
I'll keep digging at it, and am willing and able to help out and test any
patches, this is pretty high priority for me to get working again.
> StackOverflowError for hash join with round robin
> -------------------------------------------------
>
> Key: PHOENIX-2074
> URL: https://issues.apache.org/jira/browse/PHOENIX-2074
> Project: Phoenix
> Issue Type: Bug
> Reporter: James Taylor
> Attachments: PHOENIX-2074.patch
>
>
> EVENTS Table has id,article, and more columns. Id is the primay key
> MAPPING Table has id,article,category columns. Id is the primay key
> There is index on article column of both the tables.
> Below is the query.
> select count(MAPPING.article) as cnt,MAPPING.category from EVENTS
> join
> MAPPING on MAPPING.article = EVENTS.article
> group by category order by cnt ;
> Here's the stack trace:
> {code}
> Error: Encountered exception in sub plan [0] execution. (state=,code=0)
> java.sql.SQLException: Encountered exception in sub plan [0] execution.
> at
> org.apache.phoenix.execute.HashJoinPlan.iterator(HashJoinPlan.java:156)
> at
> org.apache.phoenix.jdbc.PhoenixStatement$1.call(PhoenixStatement.java:251)
> at
> org.apache.phoenix.jdbc.PhoenixStatement$1.call(PhoenixStatement.java:241)
> at org.apache.phoenix.call.CallRunner.run(CallRunner.java:53)
> at
> org.apache.phoenix.jdbc.PhoenixStatement.executeQuery(PhoenixStatement.java:240)
> at
> org.apache.phoenix.jdbc.PhoenixStatement.execute(PhoenixStatement.java:1250)
> at sqlline.Commands.execute(Commands.java:822)
> at sqlline.Commands.sql(Commands.java:732)
> at sqlline.SqlLine.dispatch(SqlLine.java:808)
> at sqlline.SqlLine.begin(SqlLine.java:681)
> at sqlline.SqlLine.start(SqlLine.java:398)
> at sqlline.SqlLine.main(SqlLine.java:292)
> Caused by: java.lang.StackOverflowError
> at
> org.apache.phoenix.iterate.RoundRobinResultIterator$RoundRobinIterator.close(RoundRobinResultIterator.java:298)
> at
> org.apache.phoenix.iterate.RoundRobinResultIterator$RoundRobinIterator.close(RoundRobinResultIterator.java:298)
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
