@pankajgaikwad  while it is wonderful that you have shared your InvokeHTTP processor configurations, I don't think enough information has been provided to provide assistance here.  All we know from you post is that some Post rest-api call was made against some service endpoint using some URL, to send form data of some content-type to which an illegal state exception was thrown. Sharing details about your use case is always helpful.  What is the target endpoint service (Polarion? which community members may not be familiar with)?  What is the full rest-api call you are trying to make? Were you able to successfully make that same rest-api call via curl local to the NiFi server? What is the structure of your NiFi FlowFile (what is in the FlowFile's content and what are the FlowFile's attributes when the FlowFile reaches the InvokeHTTOP processor)? What are the complete configurations of the invokeHTTP processor? (Some property values are cut-off in your images) What documentation are you following for this rest-api call? As far as the exception goes... Was the "java.lang.IllegalStateException: closed" accompanied by any stack trace in the the nifi-app.log? What was logged within the target service when this post request was made? I see you shared your full dataflow in another community post: An example of the original file you are obtaining via "GetFile", what attributes you are adding to the FlowFile, and how you are modifying that content before the InvokeHTTP may also be helpful here. Sharing the additional input and details may make it possible for someone in the community to provide you with some suggestions and solutions. Thank you, Matt ... View more

@HiAnil  HDF 3.5 release is based off Apache NiFi 1.12 and was released more then 5 years ago.   It was End-Of-Life as of April 2023.   NiFi-Registry service in HDF 3.5.2 only lists PostgreSQL 9.5+, 10.x, and 11.x as tested versions. I can tell you that HDF 3.5.2 has never been tested or verified against Postgres 14 or 15 and suspect there could likely be incompatibility issues.  I would suggest testing this your self before upgrading in any production environment. Keep in mind that using such an old release exposes you to CVEs addressed in the many releases put out since HDF 3.5.2.   Additional the product has had many improvements and new features added over the years. Please help our community thrive. If you found any of the suggestions/solutions provided helped you with solving your issue or answering your question, please take a moment to login and click "Accept as Solution" on one or more of them that helped. Thank you, Matt ... View more

@salahevops  I suspect you are not running Apache NiFi older then release 1.21? If so, you may be encountering this issue addressed through an improvement: https://issues.apache.org/jira/browse/NIFI-4890 Azure AD lets a lifetime on the client issued token.  That is likely 30 minutes. The token can be refreshed, but NiFi OIDC in older version does not have the ability to do the background refresh. Further improvements where added in NiFi 2.0 to add the refresh configuration timer: https://issues.apache.org/jira/browse/NIFI-12135 Please help our community thrive. If you found any of the suggestions/solutions provided helped you with solving your issue or answering your question, please take a moment to login and click "Accept as Solution" on one or more of them that helped. Thank you, Matt ... View more

@vg27  If you have a support contract with Cloudera, you could open a support case where someone could connect directly with you and assist you through your many issues. ------ 1. As i have shared before, the Single-User providers are not designed with the intent of use in a NiFi clustered environment.  They should only be used for standalone NiFi evaluation purposes.  Once you start to get in to the more involved cluster based deployments, you need to use different providers for authentication and authorization.  When using the single-user-provider for authentication, each node can create different credentials which will not work in a cluster environment.    For login based authentication, you should be using LDAP/AD (ldap-provider) or Kerberos (Kerberos-provider).  For authorization, you should be using the managed authorizer.  ------ 2. Are you still using your own generated keystore and truststore with your own created private and public certificates?  Using the NiFi auto-generated keystore and truststore will also not support clustering well as each node will not have a common certificate authority.  ----- 3. The "org.apache.zookeeper.KeeperException$ConnectionLossException: KeeperErrorCode = ConnectionLoss" exception is an issue with with Zookeeper (ZK) Quorum.  This error can happen if both you nodes are not fully up at time of exception and may also happen because you do not have proper quorum with your ZK.  Quorum consists of and odd number of ZK hosts with min 3.  Strongly encourage the use of an external ZK since anytime one of your nodes goes down, you'll lose access to both nodes. ----- 4. You are using an external https Load Balancer (LB) which means that sticky sessions (session affinity) must be setup since the user token issues when you login is only valid for use with the node that issued it.  So if your LB directs you to node 1 that presents you with login UI, you enter credentials obtaining a user token from node 1, and your LB then redirects to node 2 to load UI, it will fail authentication on node 2 because the request includes the token only good for node 1.  ----- 5. I see you are using a mix of hostnames and IP addresses in your NiFi configurations, so make sure that the node certificates include both as SAN entries to avoid issues. -----     Please help our community thrive. If you found any of the suggestions/solutions provided helped you with solving your issue or answering your question, please take a moment to login and click "Accept as Solution" on one or more of them that helped. Thank you, Matt ... View more