- Refactored NiFi Runtime class to read application properties for NAR loading and then load application properties using Framework NAR
- Removed bootstrap Class Loader from runtime initialization
- Removed nifi-properties-loader from nifi-bootstrap
- Jackson from 2.17.2 to 2.18.0
- zstd-jni from 1.5.6-5 to 1.5.6-6
- gson from 2.10.1 to 2.11.0
- okio from 3.9.0 to 3.9.1
- Fabric8 Kubernetes from 6.13.1 to 6.13.4
- Netty from 4.1.113.Final to 4.1.114.Final
- swagger-annotations from 2.2.23 to 2.2.24
- avro from 1.11.3 to 1.11.4
- commons-lang3 from 3.16.0 to 3.17.0
- log4j from 2.24.0 to 2.24.1
- Jetty from 12.0.13 to 12.0.14
- junit-bom from 5.10.3 to 5.10.4
- junit-platform-commons from 1.10.3 to 1.10.4
- mockito from 5.13.0 to 5.14.1
- testcontainers from 1.20.1 to 1.20.2
- snakeyaml from 2.2 to 2.3
- AWS SDK v2 from 2.28.4 to 2.28.13
- flyway from 10.18.0 to 10.18.2
- jline from 3.26.3 to 3.27.0
- neo4j driver from 5.24.0 to 5.25.0
- maxmind from 3.1.0 to 3.1.1
- geoip2 from 4.2.0 to 4.2.1
- amqp-client from 5.21.0 to 5.22.0
- commons-csv from 1.11.0 to 1.12.0
- splunk from 1.9.4 to 1.9.5
- lucene from 9.11.1 to 9.12.0
- Google BOM from 26.46.0 to 26.47.0
- Azure BOM from 1.2.26 to 1.2.28
- msal4j from 1.16.1 to 1.17.1
- json-schema-validator from 1.5.1 to 1.5.2
- checker-qual from 3.45.0 to 3.47.0
- checkstyle from 10.16.0 to 10.18.2
Signed-off-by: David Handermann <exceptionfactory@apache.org>
* [NIFI-13821] add Max Concurrent Tasks and Stateless Flow Timeout form fields when PG Execute Engine is STATELESS
* make fields conditionally required
This closes#9328
- Only synchronize flow with the coordinator after getting cluster response
This closes#9317
Signed-off-by: David Handermann <exceptionfactory@apache.org>
* NIFI-13721: Migrating additional details to markdown.
* NIFI-13721: Rebuilding documentation to generate mark up on the client.
* NIFI-13721: Removing old documentation approach.
* NIFI-13721: Rebase main and final fixing button colors.
* NIFI-13721: Fixing rat issues.
* NIFI-13721: Removing generated comonent documentation.
- Removing links and images from additional details markdown.
* NIFI-13721: Recreating package lock following rebase.
* NIFI-13721: Updating documentation for M3.
* NIFI-13721: Fixing broken test.
* NIFI-13721: Addressing review feedback.
* NIFI-13721: Addressing review feedback.
* NIFI-13721: Ensuring the selected item is visible after routing.
* NIFI-13721: Fixing reportingTaskTypes in finding extensions for a given type.
* NIFI-13721: Updating the see also component to handle when the referenced extension is not installed.
* NIFI-13721: Fixing expansion panel header styles.
- Ensure NARs are loaded in parent-first order during restart
- Allow replacing a NAR when other NARs are dependent on it
- Improvements to catch Throwable in several places to handle NoClassDefFoundError
Signed-off-by: David Handermann <exceptionfactory@apache.org>
This closes#9314.
- Upgraded NAR Maven Plugin from 2.0.0 to 2.1.0
- Removed nifi-api module in favor of released version
Signed-off-by: Joseph Witt <joewitt@apache.org>
- Do not reset validation status for a component if it is not in a state where validation will be performed
Signed-off-by: David Handermann <exceptionfactory@apache.org>
- Return the remote Response Stream in the Replicated Response for unknown or large content length values
- Buffered smaller responses
- Addressed code warnings in ThreadPoolRequestReplicator
Ensured that the RestartWithDifferentPort test does not allow instance reuse, since it changes the NiFi port, which can cause subsequent tests to fail if they use that same instance.
Signed-off-by: David Handermann <exceptionfactory@apache.org>
When inheriting Registry Clients from a cluster's flow, any missing registry clients are removed at the end instead of the beginning of the synchronization logic. This allows those registry clients to still be referenced while performing synchronization. Also found that if a Processor is missing from cluster's flow but is running in local flow, on startup we get an error indicating that the processor does not belong to the associated process group so fixed that in tandem. Finally, noticed while verifying the fix that we check if the proposed flow is empty with a null check instead of using the isFlowEmpty() method - this could result in inheriting an empty flow from cluster even when a flow is loaded locally.
Signed-off-by: David Handermann <exceptionfactory@apache.org>
David Handermann
Scott Aslan
Mike Moser
Matt Gilman
Vincenzo Abate
Pierre Villard
Peter Turcsanyi
dan-s1
Rajmund Takacs
Matt Burgess
Joe Gresock
lehelb
Rob Fellows
Ferenc Erdei
Ferenc Kis
Bryan Bende
dependabot[bot]
Bob Paulin
Mark Payne
Peter Kedvessy