Trigger Feed is not working because there is no enqueued and dequeued message in ActiveMQ after first feed job which is triggering another job is finished successively.
But more weird thing is above situation occurs in some part of several environments.
I tested it with same Template and Feed in 4 different environments.
But it succeeded in 2 of 4 environments and others failed.
I couldn't find any significant differences between properties of 4 environments.
I checked properties of Kylo and NiFi, controller services of NiFi.
These are brief specifications of my environments.
Environment 1 (Success) : CentOS 7/ HDP2.6.0.3-8/ Kylo0.8.1/clustered NiFi v1.0.0 on 2 nodes
Environment 2 (Success : Oracle Linux 7.2/ HDP2.6.0.3-8/ Kylo0.8.1/ NiFi v1.0.0 (It had been clustered, not now)
Environment 3 (Fail) : CentOS 7/ HDP2.6.0.3-8/ Kylo0.8.1/NiFi v1.0.0
Environment 4 (Fail) : CentOS 7/ HDP2.6.0.3-8/ Kylo0.8.1/NiFi v1.0.0
This is the log messages related to ActiveMQ and 4 environments are same.
2017-07-11 19:30:11 INFO main:ActiveMqConfig:102 - Connecting to ActiveMQ anonymously
2017-07-11 19:30:11 INFO main:ActiveMqConfig:72 - Setup ActiveMQ ConnectionFactory for tcp://localhost:61616
Additionally I can see running status of feed job in Kylo Ops Manager. NiFiKyloProvenanceReportingTask is no problem. And first feed job which is triggering another job is also finished successively.
related Q&A : https://groups.google.com/forum/#!topic/kylo-community/zbae_II1x9A
I tried to set "security.entity.access.controlled" property as comment or false, it is same with above situation.
Only enable entity-level access control, I can use TriggerFeed.