Ghanashyam created NIFI-7382:
--------------------------------
Summary: List File isn't stopped when migrated from Single node to
Clustered NiFi setup
Key: NIFI-7382
URL: https://issues.apache.org/jira/browse/NIFI-7382
Project: Apache NiFi
Issue Type: Bug
Components: Configuration, Configuration Management
Affects Versions: 1.9.2
Environment: On windows, Linux. Not tried in mac
Reporter: Ghanashyam
Attachments: NiFi-ListFile-issue.JPG, NiFi-cluster-issue.JPG,
NiFi-exceptionState-issue.JPG
NiFi was 1st deployed in single node (node06) and a flow was created like shown
below
ListFile --> FetchFile --> UpdateAttribute --> PutHDFS -->..
The flow works well in single node and was versioned using nifi registry. Once
registered, NiFi was scaled Up with 2 more nodes (node04 & node05) to form a
cluster of 3 nodes. The same version of the flow from the nifi registry was
copied on both of the new nodes.
Issue:
The primary is still the original node and the new nodes are acting like
secondary. The new nodes however does not have the directory structure which
the ListFile processor used to list the files, hence there is an exception
state in the processor.
If I wish to change the scheduling state from all nodes to primary nodes it is
not allowing this operation. It is not allowing any modifications in this state
of the processor. The error is pops up says the processor is not stopped and
is unable to fulfill the request. This is mis leading and has to be corrected
with proper statements.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
