Deepak Reddy Chirthani created NIFI-9295:
--------------------------------------------

             Summary: Issue with PutMongo(and PutMongoRecord)
                 Key: NIFI-9295
                 URL: https://issues.apache.org/jira/browse/NIFI-9295
             Project: Apache NiFi
          Issue Type: Improvement
          Components: Configuration
    Affects Versions: 1.11.4
            Reporter: Deepak Reddy Chirthani
         Attachments: image-2021-10-12-12-19-08-085.png

We are having an issue with PutMongoRecord Processor(same for PutMongo as well) 
which is routing the flowfiles to the failure relationship when the Mongo 
database(or the server hosting the db) is down. We have a set-up which looks 
like below on a high-level overview.

!image-2021-10-12-12-19-08-085.png!

As you can clearly see, all the failure relationships of different processors 
in the dataflow are connected to an Error Handling Process Group. Now, when the 
Mongodb is down, the PutMongoRecord processor is routing all the flowfiles to 
the failure relationship. Now, I know that a processor could route a flowfile 
to the failure relationship for multitude of reasons and I believe 
PutMongoRecord routing the flowfiles to failure when the database is down is 
not ideal for our scenario. Our dataflow is real-time streaming and when mongo 
database is down and flowfiles get routed to failure relationship, we get 
hundreds of emails because of the PutEmail processor in the Error Handling 
Process Group. We configured the Error Handling process group to check if we 
get any erroneous data which is against our schema or whatever maybe the reason.
 
 Is there any logic you guys help us in this scenario?
 
 I also observed that when we load the data to a RDBMS, lets say SQL Server, 
using the PutSql processor, it doesn’t route the flowfiles to the failure 
relationship when the sql server is down. I remember the flowfiles were held up 
in the connector above it until the db is up and running. I also noticed that 
PutSql has retry and PutMongo(and Record) doesn’t have any retry.
 
 Thanks in Advance. Please comment if you guys have any questions



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to