Transaction is already in progress issue

[#1]

I'm getting the "Transaction already in progress error message" when attempting to do anything in the NiFi UI. Adding a user, processor, label, etc. The only thing the UI lets me do is manage my nodes in the cluster via the cluster tab.

Some things to note:

• Version: 10.11.0
• My nifi.web.https.host value in nifi.properties lists the same value on all three nodes. (Example: user-nifi) Instead of (user-nifi1, user-nifi2, user-nifi3).
• My nifi.web.proxy.host value lists all three nodes in the instances, in addition to localhost

[#2]

@Eristocrat 

The "nifi.web.https.host" value on each node should be set to the unique hostname assigned to each of your NiFi nodes.  Or are you saying that all the three instances of NiFi in your cluster are all installed on the same server/host?   You should not have all your NiFi cluster nodes trying to bind to the same host unless they are all installed on sam host which would be extremely unusual and not recommended.

is useful when the client to NiFi presents a different hostname than what was specified in the browser URL.
For example, you enter https://nifi-proxy:443/nifi and that directs to a proxy that then routes to one of the NiFi cluster host urls https://nifi-node1:443/nifi.  What happens here is NiFi receives a request for connecting to nifi-proxy, but nifi-node1 is not nifi-proxy so it fails.  That is because it looks like someone made a request to access server nifi-proxy, but that request landed on nifi-node1.  So adding nifi-proxy to the nifi.web.proxy.host list tells NiFi to accept request on nifi-node1 that were destined for nifi-proxy.

If you found this response assisted with your query, please take a moment to login and click on "Accept as Solution" below this post.

Thank you,

Matt

[#3]

Currently the nifi.web.https.host is the same across three separate IPs/DNs as this was the solution I was given by a coworker when I was unable to access the UI. Switching them back to their unique DN causes the UI to be unreachable. Right now I'm looking through logs for a possible red herring. What is generally better, INFO or DEBUG modes for logging?

[#4]

@Eristocrat 

@Eristocrat 

Logging varies by class.  Some classes don't produce anything beyond INFO level, while others will produce even TRACE level logging.  With TRACE level being the highest level, you will get all possible logging by using it.

Do you have a screenshot of the exception you are seeing?  At time of exception, did you find anything logged in the NiFi user and/or app logs?

Thanks,

Matt