https://www.datainmotion.dev/2020/08/deleting-schemas-from-cloudera-schema.html ... View more

retry:  A FlowFile is routed to this relationship if the database cannot be updated but attempting the operation again may succeed   I think that depends on the JDBC driver and the database server which they choose to implement.   I have only ever seen failures for JDBC related things.   The code looks like it depends on what SQL Server returns through the JDBC   Have you tried PutDatabaseRecord https://github.com/apache/nifi/blob/c396927299586b896df4ebc745793b4c451f3898/nifi-nar-bundles/nifi-standard-bundle/nifi-standard-processors/src/main/java/org/apache/nifi/processors/standard/PutSQL.java   https://github.com/apache/nifi/blob/c396927299586b896df4ebc745793b4c451f3898/nifi-nar-bundles/nifi-standard-bundle/nifi-standard-processors/src/test/java/org/apache/nifi/processors/standard/TestPutSQL.java   It seems their JDBC driver may only have retry for connection issue and that would happen in the connection pool and not in the PutSQL   https://docs.microsoft.com/en-us/sql/connect/jdbc/jdbc-driver-support-for-high-availability-disaster-recovery?view=sql-server-ver15   I don't think you can work with RetryFlowFile and PutSQL with SQL Server with the generic JDBC driver.   There may be a commercial JDBC driver that has that functionality.   https://media.datadirect.com/download/docs/jdbc/alljdbc/index.html#page/jdbcconnect/Specifying_Connection_Retry_7.html   If this is a critical support issue, please open a ticket with your cloudera support account. ... View more

looks like some Hadoop interfacing issues, open a ticket with Cloudera support.   https://community.cloudera.com/t5/Support-Questions/DataXceiver-error-processing-unknown-operation-java-io/td-p/164048 ... View more

That is really odd, I recommend you open a ticket with your Cloudera support.   I think having an engineer walk through your environment may be the only way to find such a complex issue.   This one is really odd.   I haven't seen this occur unless there was an issue with security or HDFS.   Or missing / broken file path or file permissions https://community.cloudera.com/t5/Support-Questions/putHDFS-processor-errors-in-Nifi/td-p/191407   You could try PutOrc or PutFile.    Can that NiFi reach any other Kerberos protected resources like HBase or Hive?   Dual NIC issue (networking) https://community.cloudera.com/t5/Support-Questions/Nifi-putHDFS-throwing-replication-error-for-single-node/td-p/199784   Invalid Private Address https://community.cloudera.com/t5/Support-Questions/Nifi-Processor-PutHDFS-not-working-to-write-files-on-HDFS/td-p/237546   bootstrap.conf issue  java.arg.101=-Djavax.security.auth.useSubjectCredsOnly=true https://community.cloudera.com/t5/Support-Questions/Nifi-PutHDFS-Processor-stuck-and-not-able-to-release-item/td-p/229780   Keypath https://community.cloudera.com/t5/Support-Questions/How-to-connect-to-Kerberized-HDP-cluster-from-separate/td-p/152730   https://risdenk.github.io/2018/03/15/hdf-apache-nifi-kerberos-errors-usesubjectcredsonly.html   Login time https://community.cloudera.com/t5/Support-Questions/PutHDFS-Impersonation-error/td-p/111628   https://community.cloudera.com/t5/Support-Questions/NiFi-PutHDFS-Login-Failure-using-kerberos/td-p/195595/page/2 https://community.cloudera.com/t5/Support-Questions/how-to-use-apache-Nifi-on-kerberized-HDP-Cluster-nodes/td-p/116503 ... View more

Most people doing give root accounts HDFS accounts or kerberos accounts.   Depends on the environment.   Just trying to see what is different between environments.   If it's not user, permissions, firewall, version of NiFi, that's weird.   I have seen some odd setups of Linux where security intercepts somethings or prevents root users from doing certain things.   i have seen devops tools change permissions automatically in the background.   You can change PutHDFS to Primary node to have just one write to it to narrow down which machine in the cluster may have an issue.     ... View more

Could be local security or file permissions or IPtables / Firewall between NiFi cluster and HDFS.   https://www.edureka.co/community/30977/hadoop-hdfs-exception-in-createblockoutputstream   Could be crypto libraries installed?   https://community.cloudera.com/t5/Support-Questions/NiFi-PutHDFS-Writing-Zero-Bytes-appears-to-be-crypto-related/td-p/115493   If this NiFi installed via CDF/CFM Cloudera Manager?   If so, please open a support ticket. ... View more

What kerberos ID are you logged into on your PC vs in the NiFi cluster.   Perhaps the NiFi kerberos ID does not have WRITE permissions to that directory or you are not logged into Kerberos on NiFi cluster.   People usually don't have the same kerberos credentials on their PC as on their cluster.    Should be a service user not a personal user. ... View more

===  There are a number of possible of causes for this. The NameNode may be overloaded. Check the logs for messages that say "discarding calls..." There may not be enough (any) DataNode nodes running for the data to be written. Again, check the logs. Every DataNode on which the blocks were stored might be down (or not connected to the NameNode; it is impossible to distinguish the two).   Looks like HDFS issues, can you post from Hue or HDFS command line   https://cwiki.apache.org/confluence/display/HADOOP2/TroubleShooting   ... View more

Can you post configuration information, more logs, plus some details on:   NiFi version HDFS version Encrypted file system? Cloud? OS version JDK version / JVM version CDH/HDP/CDP version PutHDFS Settings What kind of data?  Example data ... View more

That's good news.   That's a tricky one.   Glad things are working for you. ... View more