[
https://issues.apache.org/jira/browse/PHOENIX-6736?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Istvan Toth updated PHOENIX-6736:
---------------------------------
Description:
Running IndexTool with the "-pr" option results in an NPE
{noformat}
hbase org.apache.phoenix.mapreduce.index.IndexTool
-Dhbase.client.scanner.timeout.period=3600000 -Dhbase.rpc.timeout=3600000
-Dmapreduce.task.timeout=57600000 --data-table BUBU -pr -op /tmp/MRtest3
...
java.lang.NullPointerException
at
org.apache.phoenix.mapreduce.index.IndexTool.submitIndexToolJob(IndexTool.java:861)
at org.apache.phoenix.mapreduce.index.IndexTool.run(IndexTool.java:828)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:81)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:95)
at
org.apache.phoenix.mapreduce.index.IndexTool.main(IndexTool.java:1235){noformat}
The stack trace is from a slightly older 5.1 version, but it is thrown at the
{code:java}
pDataTable.getPhysicalName().getString() : {code}
line.
The table I tested with has the new indexing.
* Do we even support partial rebuilds for the new indexing ? I strongly
suspect not, in which case we should have refused the command earlier
* I rolled back to the old indexing (with IndexUpgradeTool), and I got the
same error, so this looks like a regression for old style tables.
was:
Running IndexTool with the "-pr" option results in an NPE
{noformat}
hbase org.apache.phoenix.mapreduce.index.IndexTool
-Dhbase.client.scanner.timeout.period=3600000 -Dhbase.rpc.timeout=3600000
-Dmapreduce.task.timeout=57600000 --data-table BUBU -pr -op /tmp/MRtest3
...
java.lang.NullPointerException
at
org.apache.phoenix.mapreduce.index.IndexTool.submitIndexToolJob(IndexTool.java:861)
at org.apache.phoenix.mapreduce.index.IndexTool.run(IndexTool.java:828)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:81)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:95)
at
org.apache.phoenix.mapreduce.index.IndexTool.main(IndexTool.java:1235){noformat}
The stack trace is from a slightly older 5.1 version, but it is thrown at the
{code:java}
pDataTable.getPhysicalName().getString() : {code}
line.
The table I tested with has the new indexing.
* Do we even support partial rebuilds for the new indexing ? I strongly suspect
not, in which case we should have refused the command earlier
* Do we get the same error on a table with the old indexing ? If yes, then
that's a backwards compatibility regression that we should fix.
> IndexTool partial rebuild throws NPE
> ------------------------------------
>
> Key: PHOENIX-6736
> URL: https://issues.apache.org/jira/browse/PHOENIX-6736
> Project: Phoenix
> Issue Type: Bug
> Components: core
> Affects Versions: 5.2.0
> Reporter: Istvan Toth
> Priority: Major
>
> Running IndexTool with the "-pr" option results in an NPE
> {noformat}
> hbase org.apache.phoenix.mapreduce.index.IndexTool
> -Dhbase.client.scanner.timeout.period=3600000 -Dhbase.rpc.timeout=3600000
> -Dmapreduce.task.timeout=57600000 --data-table BUBU -pr -op /tmp/MRtest3
> ...
> java.lang.NullPointerException
> at
> org.apache.phoenix.mapreduce.index.IndexTool.submitIndexToolJob(IndexTool.java:861)
> at org.apache.phoenix.mapreduce.index.IndexTool.run(IndexTool.java:828)
> at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:81)
> at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:95)
> at
> org.apache.phoenix.mapreduce.index.IndexTool.main(IndexTool.java:1235){noformat}
> The stack trace is from a slightly older 5.1 version, but it is thrown at the
> {code:java}
> pDataTable.getPhysicalName().getString() : {code}
> line.
> The table I tested with has the new indexing.
> * Do we even support partial rebuilds for the new indexing ? I strongly
> suspect not, in which case we should have refused the command earlier
> * I rolled back to the old indexing (with IndexUpgradeTool), and I got the
> same error, so this looks like a regression for old style tables.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
