You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@oozie.apache.org by pu...@apache.org on 2017/01/31 06:55:13 UTC

oozie git commit: OOZIE-2782 Input logic wait documentation is confusing

Repository: oozie
Updated Branches:
  refs/heads/master 6fa03855c -> c26615f9f


OOZIE-2782 Input logic wait documentation is confusing


Project: http://git-wip-us.apache.org/repos/asf/oozie/repo
Commit: http://git-wip-us.apache.org/repos/asf/oozie/commit/c26615f9
Tree: http://git-wip-us.apache.org/repos/asf/oozie/tree/c26615f9
Diff: http://git-wip-us.apache.org/repos/asf/oozie/diff/c26615f9

Branch: refs/heads/master
Commit: c26615f9ff7ebab0dd9d66ac66d5d2c12fd3145b
Parents: 6fa0385
Author: puru <pu...@gmail.com>
Authored: Mon Jan 30 22:54:59 2017 -0800
Committer: puru <pu...@gmail.com>
Committed: Mon Jan 30 22:54:59 2017 -0800

----------------------------------------------------------------------
 .../src/site/twiki/CoordinatorFunctionalSpec.twiki | 17 +++++++++++++++--
 release-log.txt                                    |  1 +
 2 files changed, 16 insertions(+), 2 deletions(-)
----------------------------------------------------------------------


http://git-wip-us.apache.org/repos/asf/oozie/blob/c26615f9/docs/src/site/twiki/CoordinatorFunctionalSpec.twiki
----------------------------------------------------------------------
diff --git a/docs/src/site/twiki/CoordinatorFunctionalSpec.twiki b/docs/src/site/twiki/CoordinatorFunctionalSpec.twiki
index 744f5d9..4b21ad9 100644
--- a/docs/src/site/twiki/CoordinatorFunctionalSpec.twiki
+++ b/docs/src/site/twiki/CoordinatorFunctionalSpec.twiki
@@ -3461,9 +3461,10 @@ Supported operators are OR, AND, COMBINE. OR and AND operators are nested, one c
    * COMBINE :  With combine, instances of A and B can be interleaved to get the final "combined" set of total instances. All datasets in combine should have the same range defined with the current EL function. Combine does not support latest and future EL functions. Combine cannot also be nested.
 
 Additional options
-   * *%BLUE% MIN: %ENDCOLOR%* Minimum number of input instances that should be available.
-   * *%BLUE% WAIT (in minutes): %ENDCOLOR%* If all dependencies are not met, and MIN dependencies are met,  then Oozie will keep on waiting for more instances till wait time elapses or all dependent data are available.
+   * *%BLUE% MIN: %ENDCOLOR%* Minimum number of input instances that should be available. This can be used in cases where inputs are optional and the processing can be done on a subset of input instances.
+   * %BLUE% WAIT (in minutes): %ENDCOLOR% Wait is used to specify the amount of time to continue checking for availability of instances of a particular dataset before moving on to the next dataset defined in a OR condition. The wait time is calculated from the beginning of the nominal time of the action or the action creation time whichever is later. The main purpose of this is to give preference to the primary datasource before checking the secondary datasource.
 
+Wait when used with min option has a totally different purpose. It is used to specify the additional amount of time to wait and check for more instances after the required minimum set of instances become available. Any additional instances that become available during the wait time are then included.
 
 The conditional logic can be specified using the <input-logic> tag in the coordinator.xml using the [[CoordinatorFunctionalSpec#Oozie_Coordinator_Schema_0.5][Oozie Coordinator Schema 0.5]] and above. If not specified, the default behavior of "AND" of all defined input dependencies is applied.
 
@@ -3484,6 +3485,18 @@ With above expression one can specify the dataset as AorB. Action will start run
 
 *%GREEN% Example: %ENDCOLOR%*:
 <verbatim>
+<input-logic>
+    <or name="AorBorC">
+        <data-in dataset="A" wait = "60"/>
+        <data-in dataset="B" wait = "90"/>
+        <data-in dataset="C"/>
+    </or>
+</input-logic>
+</verbatim>
+With the above expression, it will wait for 60 mins from the nominal time of the action or the action creation time whichever is later for all the instances of dataset A to be available. If it is not available in 60 minutes, then it will start checking for instances of dataset B. If instances of B are not available in another 30 minutes, then it will start checking for dataset C.
+
+*%GREEN% Example: %ENDCOLOR%*:
+<verbatim>
 <datasets>
        <dataset name="dataset_a" frequency="${coord:minutes(20)}" initial-instance="2010-01-01T00:00Z" timezone="UTC">
             <uri-template>${nameNode}/user/${coord:user()}/${examplesRoot}/input-data/rawLogs/${YEAR}/${MONTH}/${DAY}/${HOUR}/${MINUTE}</uri-template>

http://git-wip-us.apache.org/repos/asf/oozie/blob/c26615f9/release-log.txt
----------------------------------------------------------------------
diff --git a/release-log.txt b/release-log.txt
index 2b806fc..f7fd833 100644
--- a/release-log.txt
+++ b/release-log.txt
@@ -1,5 +1,6 @@
 -- Oozie 4.4.0 release (trunk - unreleased)
 
+OOZIE-2782 Input logic wait documentation is confusing (puru)
 OOZIE-2781 HCat partition available notification is not sent to coordinator actions if coordinator job is using a different hostname (cname, IP address, etc. ) for HCat URL (puru)
 OOZIE-2770 Show missing dependencies for coord actions (puru)
 OOZIE-2630 Oozie Coordinator EL Functions to get first day of the week/month (satishsaley)