devinbost opened a new pull request #4012: Adding upsert functionality
URL: https://github.com/apache/pulsar/pull/4012
 
 
   ### Motivation
   Upsert functionality is intended to simplify continuous deployment 
processes. This simplification is especially desirable because there is no 
Python Admin API available that provides typed objects. Rather, users are 
required to build workflows with the Admin CLI, which returns strings that can 
be messy to handle. Upsert helps solve this complexity by allowing Pulsar to 
handle behavior according to whether the component has already been created or 
not. 
   
   ### Motivation (with future in mind)
   Upsert is a dependency of BulkUpsert (to be added), which will greatly 
simplify seamless continuous deployments by allowing Pulsar to rapidly update 
Functions, Sinks, and Sources (and leverage parallelization). BulkUpsert will 
enable automated deployments to handle all but deletes on components that are 
no longer used. (For that, functionality will need to be added to allow the 
user to easily compare their expected component tree with the component tree in 
Pulsar to identify components that need to be deleted.)
   
   ### Modifications
   My contributions add Upsert functionality for Functions, Sinks, and Sources 
to the REST API and to the Admin CLI. Upsert conditionally combines the 
functionality of Create and Update operations. If the component already exists 
in Pulsar, Upsert updates the component; if the component doesn't already exist 
in Pulsar, Upsert creates the component.
   
   Commits 29e2a66 and 7749419 were just to get the latest changes from the 
Pulsar repo. 
   My primary code is in commit 441ba03, and the unit tests are in commit 
441ba03. 
   
   ### Verifying this change
   This commit is largely covered already by unit tests for registration and 
updates for Functions, Sinks, and Sources. However, additional unit tests have 
been added for Upsert specifically. Tests were added to:
   
   - org.apache.pulsar.functions.worker/rest.api/FunctionApiV3ResourceTest.java
   - org.apache.pulsar.functions.worker/rest.api/SinkApiV3ResourceTest.java
   - org.apache.pulsar.functions.worker/rest.api/SourceApiV3ResourceTest.java
   - org.apache.pulsar.functions.worker/FunctionMetaDataManagerTest.java
   
   When running the build tests, I did notice these **unrelated** test failures:
   1. 
   > [ERROR] Tests run: 2, Failures: 1, Errors: 0, Skipped: 1, Time elapsed: 
13.506 s <<< FAILURE! - in org.apache.pulsar.broker.service.RackAwareTest
   > [ERROR] testPlacement(org.apache.pulsar.broker.service.RackAwareTest)  
Time elapsed: 1.085 s  <<< FAILURE!
   > java.lang.AssertionError: first bookie in rack 0 not included in ensemble 
expected [true] but found [false]
   2. 
   
   > [ERROR] Tests run: 13, Failures: 1, Errors: 0, Skipped: 1, Time elapsed: 
59.026 s <<< FAILURE! - in org.apache.pulsar.io.PulsarFunctionE2ETest
   > [ERROR] 
testPulsarSinkStatsWithUrl(org.apache.pulsar.io.PulsarFunctionE2ETest)  Time 
elapsed: 0.614 s  <<< FAILURE!
   > java.lang.NullPointerException
   >    at 
org.apache.pulsar.io.PulsarFunctionE2ETest.testPulsarSinkStats(PulsarFunctionE2ETest.java:527)
   >    at 
org.apache.pulsar.io.PulsarFunctionE2ETest.testPulsarSinkStatsWithUrl(PulsarFunctionE2ETest.java:708)
   
   It appears that these test failures are unrelated to any changes that I made 
to this repo. However, I ran these tests before merging in commit 29e2a66. 
(Sorry if that's confusing. I was working in a different directory.)
   
   ### This pull request affects:
     - The public API: (**yes**)
     - The rest endpoints: (**yes**)
     - The admin cli options: (**yes**)
     - Anything that affects deployment: (don't know)
   
   ### Documentation
   
     - Does this pull request introduce a new feature? (**yes**)
     - If yes, how is the feature documented? (**not documented YET**). (Please 
provide guidance on where it needs to be documented. There are 930 markdown 
(.md) files in this repo...)
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to