You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@trafodion.apache.org by db...@apache.org on 2016/03/15 21:20:05 UTC
[6/7] incubator-trafodion git commit: Incorporated review changes
from Dave Birdsall. Thanks!
Incorporated review changes from Dave Birdsall. Thanks!
Also fixed too-dumb-to-live issues in the Win-ODBC chapter in the Client
Installation Guide.
Project: http://git-wip-us.apache.org/repos/asf/incubator-trafodion/repo
Commit: http://git-wip-us.apache.org/repos/asf/incubator-trafodion/commit/6ad54f0b
Tree: http://git-wip-us.apache.org/repos/asf/incubator-trafodion/tree/6ad54f0b
Diff: http://git-wip-us.apache.org/repos/asf/incubator-trafodion/diff/6ad54f0b
Branch: refs/heads/master
Commit: 6ad54f0b25f96970828d7f934cbd6ba64fe32b1b
Parents: ea2c1a0
Author: Gunnar Tapper <gt...@esgyn.local>
Authored: Tue Mar 15 12:42:30 2016 -0600
Committer: Gunnar Tapper <gt...@esgyn.local>
Committed: Tue Mar 15 12:42:30 2016 -0600
----------------------------------------------------------------------
.../src/asciidoc/_chapters/odbc_windows.adoc | 10 +++++-----
.../src/asciidoc/_chapters/histograms.adoc | 12 +++++-------
.../src/asciidoc/_chapters/operational_controls.adoc | 12 ++++++------
.../src/asciidoc/_chapters/query_execution.adoc | 14 +++++++-------
docs/load_transform/src/asciidoc/_chapters/about.adoc | 2 +-
.../src/asciidoc/_chapters/bulk_load.adoc | 4 ++--
6 files changed, 26 insertions(+), 28 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/incubator-trafodion/blob/6ad54f0b/docs/client_install/src/asciidoc/_chapters/odbc_windows.adoc
----------------------------------------------------------------------
diff --git a/docs/client_install/src/asciidoc/_chapters/odbc_windows.adoc b/docs/client_install/src/asciidoc/_chapters/odbc_windows.adoc
index 8c20cfd..739d0e6 100644
--- a/docs/client_install/src/asciidoc/_chapters/odbc_windows.adoc
+++ b/docs/client_install/src/asciidoc/_chapters/odbc_windows.adoc
@@ -24,7 +24,7 @@
These installation instructions apply to Trafodion Release 1.3.*_n_* and subsequent updates of the product until otherwise indicated.
== Installation Package
-The installer file for the Trafodion ODBC Driver for Windows is packaged within the Trafodion clients tar file, `clients-1.3.*_n_*.tar.gz`, on the
+The installer file for the Trafodion ODBC Driver for Windows is packaged within the Trafodion clients tar file, `Trafodion Clients`, on the
Trafodion downloads site http://trafodion.incubator.apache.org/download.html. Use this installer file to install the Trafodion ODBC Driver for Windows:
[cols="25%l,25%,50%",options="header"]
@@ -73,10 +73,10 @@ NOTE: To install the driver on your PC, you must be logged on with a user ID tha
==== Task 1: Download the ODBC client installation file
1. Create a download folder on the client workstation.
-2. Open a Web browser and navigate to the Trafodion downloads site http://trafodion.incubator.apache.com/download.html
-3. Click on the `clients-1.3.*_n_*.tar.gz` link to start downloading the Trafodion clients tar file to your workstation.
-4. Place the `clients-1.3.*_n_*.tar.gz file` into the download folder.
-5. Extract the contents of the `clients-1.3.*_n_*.tar.gz` file by using an extract tool like WinZip. The package file contains the
+2. Open a Web browser and navigate to the Trafodion downloads site http://trafodion.incubator.apache.org/download.html
+3. Click on the `Trafodion Clients` link to start downloading the Trafodion clients tar file to your workstation.
+4. Place the `*.tar.gz file` into the download folder.
+5. Extract the contents of the `*.tar.gz` file by using an extract tool like WinZip. The package file contains the
`TFODBC64-1.3.*_n_*.msi` installer file, which is extracted to the clients sub-directory.
==== Task 2: Start the InstallShield wizard
http://git-wip-us.apache.org/repos/asf/incubator-trafodion/blob/6ad54f0b/docs/cqd_reference/src/asciidoc/_chapters/histograms.adoc
----------------------------------------------------------------------
diff --git a/docs/cqd_reference/src/asciidoc/_chapters/histograms.adoc b/docs/cqd_reference/src/asciidoc/_chapters/histograms.adoc
index 2126f53..7086d90 100644
--- a/docs/cqd_reference/src/asciidoc/_chapters/histograms.adoc
+++ b/docs/cqd_reference/src/asciidoc/_chapters/histograms.adoc
@@ -53,7 +53,7 @@ make the refresh more frequent.
| *Impact* | Longer histogram refresh intervals can improve compile times. However, longer refresh intervals yield more obsolete
the histograms. More obsolete histograms may result in poor performance for queries that could leverage recently updated statistics.
| *Level* | System or Service.
-| *Conflicts/Synergies* | Frequency of update statistics run either using MAINTAIN.
+| *Conflicts/Synergies* | Frequency of update statistics run using MAINTAIN.
| *Real Problem Addressed* | Not applicable.
| *Introduced In Release* | Trafodion 1.3.0.
| *Deprecated In Release* | Not applicable.
@@ -67,19 +67,18 @@ the histograms. More obsolete histograms may result in poor performance for quer
|===
| *Description* | Controls the level of missing statistics warnings that should be displayed. The warnings impacted are 6007, 6008, 6010, and 6011.
| *Values* |
-*'0'*: Display no warnings. The Update Statistics Automation Server still continues to record missing column statistics. +
+*'0'*: Display no warnings. +
+
*'1'*: Display only missing single column statistics warnings. These include 6008 and 6011. +
+
*'2'*: Display all missing single and multi-column statistics warnings for scans only. +
+
-*'3'*: Display all missing single and multi-column statistics warnings for scans and join operators only.
-The Update Statistics Automation Server continues to record missing column statistics. +
+*'3'*: Display all missing single and multi-column statistics warnings for scans and join operators only. +
+
*'4'*: Display all missing single and multi-column statistics warnings. +
+
The default value is *'4'*.
-| *Usage* | The Update Statistics Automation Server (USAS) captures these warnings and gathers appropriate statistics.
+| *Usage* |
If you do not want to see these warnings, then change the setting to '0'. If you want to track the warnings, then you have a
choice of which warnings you want to track. Each setting gives you the ability to filter the warnings seen for missing single
or multi-column statistics for join or scan operations. This controls the resulting number of warning messages. +
@@ -99,8 +98,7 @@ to get several warnings. Not all warnings may contribute to plan improvements. T
based on the search path, some of which may not even impact the plan quality. Also, the cost of gathering statistics on those columns
may not bring commensurate benefit to a large number of queries.
| *Level* | System.
-| *Conflicts/Synergies* | The USAS does capture these warnings and gathers statistics felt appropriate
-for query performance. If you are using USAS, then you may set the warning level to 1 or 0.
+| *Conflicts/Synergies* | Not applicable.
| *Real Problem Addressed* | Not applicable.
| *Introduced In Release* | Trafodion 1.3.0.
| *Deprecated In Release* | Not applicable.
http://git-wip-us.apache.org/repos/asf/incubator-trafodion/blob/6ad54f0b/docs/cqd_reference/src/asciidoc/_chapters/operational_controls.adoc
----------------------------------------------------------------------
diff --git a/docs/cqd_reference/src/asciidoc/_chapters/operational_controls.adoc b/docs/cqd_reference/src/asciidoc/_chapters/operational_controls.adoc
index 50d95a2..8d077cf 100644
--- a/docs/cqd_reference/src/asciidoc/_chapters/operational_controls.adoc
+++ b/docs/cqd_reference/src/asciidoc/_chapters/operational_controls.adoc
@@ -114,13 +114,13 @@ are cached and larger scans are not cached to avoid cache trashing.
*> 0*: The cached Hive meta-data is valid only for <value> seconds. +
+
Default: *'0'*.
-| *Usage* | Use of value -1 when the Hive tables are read-only and no updates are possible to avoid reading the
-Hive metadata when the same table is referenced in queries again. +
+| *Usage* | Use a value of -1 when the Hive tables are read-only. This allows Trafodion to avoid repeatedly reading the
+metadata when the table is referenced in multiple queries. +
+
-Use of value 0 when frequent update of Hive tables are possible. +
+Use a value of 0 when updates to the Hive tables are likely to be frequent. +
+
-Use of a positive value when when infrequent update of Hive tables can take place and the value can be set so that it is impossible
-for the Hive tables to be updated within next <value> seconds.
+Use a positive value _n_ to cause Trafodion to re-read metadata after n seconds has elapsed.
+Use this option when updates to the Hive table happen at least _n_ seconds apart.
| *Production Usage* | Consult with the Trafodion community.
| *Impact* | Compilation time.
| *Level* | Session.
@@ -200,7 +200,7 @@ If the plan object is not produced due to a text cache miss, then the plan is st
Default: *'OFF'*.
| *Usage* | When this CQD is ON during a bulk load, then any indexes on a table are maintained incrementally.
New rows are added to the base table and all the indexes in HFiles and then during LOAD COMPLETE phase all new files are moved to HBase.
-Indexes are not offline. However, it causes the index to be inconsistent with the base table if any of the new rows have the same key value as an existing.
+Indexes are not offline. However, it causes the index to be inconsistent with the base table if any of the new rows have the same key value as an existing row.
Change the default to ON only when certain that new rows do not have a conflict with existing rows in table.
| *Production Usage* | Yes.
| *Impact* | Bulk load into tables with index is faster, when the attribute is set to ON.
http://git-wip-us.apache.org/repos/asf/incubator-trafodion/blob/6ad54f0b/docs/cqd_reference/src/asciidoc/_chapters/query_execution.adoc
----------------------------------------------------------------------
diff --git a/docs/cqd_reference/src/asciidoc/_chapters/query_execution.adoc b/docs/cqd_reference/src/asciidoc/_chapters/query_execution.adoc
index c2553f1..eae1ff3 100644
--- a/docs/cqd_reference/src/asciidoc/_chapters/query_execution.adoc
+++ b/docs/cqd_reference/src/asciidoc/_chapters/query_execution.adoc
@@ -71,7 +71,7 @@ data blocks cached earlier thereby impacting the performance of the random reads
in such a case. +
+
Set the CQD <<hbase-region-server-max-heap-size,HBASE_REGION_SERVER_MAX_HEAP_SIZE>> value to reflect the amount of java heap size reserved for the region servers.
-This CQD is used by the Trafodion Optimizer to evaluate if the block cache to be turned off.
+This CQD is used by the Trafodion Optimizer to evaluate if the block cache should be turned off.
| *Production Usage* | Leave the setting to be 'SYSTEM' when HBase is configured to use LRU block cache. If needed,
you can override this settings with 'ON' or 'OFF'. +
+
@@ -162,8 +162,8 @@ if you think that you need to use this CQD.
|===
<<<
-[[hbase-rowset-vssb-opt]]
-== HBASE_ROWSET_VSSB_OPT
+[[hbase-rowset-vsbb-opt]]
+== HBASE_ROWSET_VSBB_OPT
[cols="25%h,75%"]
|===
@@ -187,8 +187,8 @@ If possible, then the query plan involves VSBB operators. The Virtual Sequential
|===
<<<
-[[hbase-rowset-vssb-size]]
-== HBASE_ROWSET_VSSB_SIZE
+[[hbase-rowset-vsbb-size]]
+== HBASE_ROWSET_VSBB_SIZE
[cols="25%h,75%"]
|===
@@ -201,7 +201,7 @@ The default value is *'1024'*.
the queue to IUD (INSERT,UPDATE,DELETE) operator is filled up in the data flow architecture of Trafodion. +
+
You can adjust the maximum size to suit your application needs and thus tune it to perform optimally.
-| *Production Usage* | Yes. You can disable this feature by setting the HBASE_ROWSET_VSSB_OPT CQD to 'OFF'.
+| *Production Usage* | Yes. You can disable this feature by setting the HBASE_ROWSET_VSBB_OPT CQD to 'OFF'.
| *Impact* | The performance of your application may be affected by setting this CQD too low.
| *Level* | Query.
| *Conflicts/Synergies* | Not applicable.
@@ -216,7 +216,7 @@ You can adjust the maximum size to suit your application needs and thus tune it
[cols="25%h,75%"]
|===
-| *Description* | Enables Trafodion to leverage the HBase small scanner optimization. This optimization reduces I/O usage with up to 66%
+| *Description* | Enables Trafodion to leverage the HBase small scanner optimization. This optimization reduces I/O usage up to 66%
and enables non-blocking reads for higher concurrency support. When a scan is known to require less than a HBASE BLOCK SIZE (default is 64K),
then enabling the HBase small scanner optimization increases performance.
| *Values* |
http://git-wip-us.apache.org/repos/asf/incubator-trafodion/blob/6ad54f0b/docs/load_transform/src/asciidoc/_chapters/about.adoc
----------------------------------------------------------------------
diff --git a/docs/load_transform/src/asciidoc/_chapters/about.adoc b/docs/load_transform/src/asciidoc/_chapters/about.adoc
index 4a1a1bf..f15fd04 100644
--- a/docs/load_transform/src/asciidoc/_chapters/about.adoc
+++ b/docs/load_transform/src/asciidoc/_chapters/about.adoc
@@ -37,7 +37,7 @@ This guide targets anyone wanting to load data into a Trafodion database.
You need to have skills in the following areas to make full use of the information in this guide:
* SQL DDL and DML.
-* Installion and configuration of Linux software.
+* Installation and configuration of Linux software.
* Trafodion administration.
* Depending on your data source, Java and/or Hadoop ecosystem usage.
http://git-wip-us.apache.org/repos/asf/incubator-trafodion/blob/6ad54f0b/docs/load_transform/src/asciidoc/_chapters/bulk_load.adoc
----------------------------------------------------------------------
diff --git a/docs/load_transform/src/asciidoc/_chapters/bulk_load.adoc b/docs/load_transform/src/asciidoc/_chapters/bulk_load.adoc
index f17f718..46593b5 100644
--- a/docs/load_transform/src/asciidoc/_chapters/bulk_load.adoc
+++ b/docs/load_transform/src/asciidoc/_chapters/bulk_load.adoc
@@ -52,8 +52,8 @@ LOAD INTO target_table SELECT * FROM source_table WHERE custkey >= 1000 ;
[[bulk-load-data-from-hdfs-files]]
== Load Data From HDFS Files
-You copy your data (local or remote) into an HDFS folder. Then, you create an external Hive table (with correct fields) that point to the HDFS folder containing the data.
-You may also specify a WHERE clause on the source data as a filter, if needed.
+You copy your data (local or remote) into an HDFS folder. Then, you create an external Hive table (with correct fields) that points
+to the HDFS folder containing the data. You may also specify a WHERE clause on the source data as a filter, if needed.
See the https://cwiki.apache.org/confluence/display/Hive/LanguageManual+DDL#LanguageManualDDL-ExternalTables[External Tables]
page on the https://cwiki.apache.org/confluence/display/Hive[Hive Wiki] for more information.