Found two more from HIVE-5522<https://issues.apache.org/jira/browse/HIVE-5522>
 (also HIVE-6098 <https://issues.apache.org/jira/browse/HIVE-6098>, Merge
Tez branch into trunk) so the current total is 51 configs that don't have
descriptions in 0.13.0:

*Release 13 *

hive.compute.splits.in.am

hive.rpc.query.plan


But these both have comments in HiveConf.java that can be used as
descriptions, although they aren't included in hive-default.xml.template.
 I missed them because I was working from the patch for
HIVE-6037<https://issues.apache.org/jira/browse/HIVE-6037> and
Navis had used the HiveConf comments for descriptions.  (That means there
could be more parameters missing from the 0.13.0 template file.)



-- Lefty


On Mon, Apr 14, 2014 at 1:53 AM, Lefty Leverenz <leftylever...@gmail.com>wrote:

> Here's a list of 49 configuration parameters in RC0 (and trunk) that don't
> have descriptions in hive-default.xml.template:
>
>
> *Release 1 or 2 *
>
> hive.exec.submitviachild
>
> hive.metastore.metadb.dir
>
> hive.jar.path
>
> hive.aux.jars.path
>
> hive.table.name
>
> hive.partition.name
>
> hive.alias
>
>
> *Release 3 *
>
> hive.cli.errors.ignore
>
>
> *Release 4 *
>
> hive.added.files.path
>
> hive.added.jars.path
>
>
> *Release 5 *
>
> hive.intermediate.compression.codec
>
> hive.intermediate.compression.type
>
> hive.added.archives.path
>
>
> *Release 6 *
>
> hive.metastore.archive.intermediate.archived
>
> hive.metastore.archive.intermediate.extracted
>
> hive.mapred.partitioner
>
> hive.exec.script.trust
>
> hive.hadoop.supports.splittable.combineinputformat
>
>
> *Release 7 *
>
> hive.lockmgr.zookeeper.default.partition.name
>
> hive.metastore.fs.handler.class
>
> hive.query.result.fileformat
>
> hive.hashtable.initialCapacity
>
> hive.hashtable.loadfactor
>
> hive.debug.localtask
>
> hive.lock.manager
>
> hive.outerjoin.supports.filters
>
> hive.semantic.analyzer.hook
>
>
> *Release 8 *
>
> hive.exec.job.debug.timeout
>
> hive.exec.tasklog.debug.timeout
>
> hive.merge.rcfile.block.level
>
> hive.merge.input.format.block.level
>
> hive.merge.current.job.has.dynamic.partitions
>
> hive.stats.collect.rawdatasize
>
>
> *Release 8.1 *
>
> hive.optimize.metadataonly
>
>
> *Release 9 *
>
>
> *Release 10 *
>
>
> *Release 11 *
>
> hive.exec.rcfile.use.sync.cache
>
> hive.stats.key.prefix                                    --- *internal*
>
>
> *Release 12 *
>
> hive.scratch.dir.permission
>
> datanucleus.fixedDatastore
>
> datanucleus.rdbms.useLegacyNativeValueStrategy
>
> hive.optimize.sampling.orderby                     --- *internal?*
>
> hive.optimize.sampling.orderby.number
>
> hive.optimize.sampling.orderby.percent
>
> hive.server2.authentication.ldap.Domain
>
> hive.server2.session.hook
>
> hive.typecheck.on.insert
>
>
> *Release 13 *
>
> hive.metastore.expression.proxy
>
> hive.txn.manager
>
> hive.stageid.rearrange
>
> hive.explain.dependency.append.tasktype
>
>
>
> What's the best way to deal with these?
>
>    1. Ignore them (or identify those that can be ignored).
>    2. Add some descriptions in Hive 0.13.0 RC1.
>    3. Deal with them after 
> HIVE-6037<https://issues.apache.org/jira/browse/HIVE-6037>gets committed.
>       - Try to cover all of them by Hive 0.14.0:
>       - Put the list in a JIRA and create a common HiveConf.java patch,
>          which can be appended until release 0.14.0 is ready.
>          - Accumulate descriptions in JIRA comments, then create a patch
>          from the comments.
>       - Deal with them as soon as possible:
>       - Put the list in an umbrella JIRA and use sub-task JIRAs to add
>          descriptions individually or in small groups.
>          4. Deal with them in the wiki, then patch HiveConf.java before
>    release 0.14.0.
>    5. [Your idea goes here.]
>
>
> -- Lefty
>

Reply via email to