[
https://issues.apache.org/jira/browse/NIFI-599?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14532551#comment-14532551
]
Joseph Witt commented on NIFI-599:
----------------------------------
Miriam,
The thought process here is that if you have a node down that you are ok with
having it down/out of the cluster then you can and should remove it from the
cluster using the cluster admin options. This is because while a node is out
of the cluster there is a whole range of things we do not allow to occur to
ensure consistency. Once that node has been removed from the cluster through
that admin option you can use provenance and the whole range of features again.
It is possible we can address this for provenance specifically because I do
share your view that it seems like this should be fine. But we do then have to
consider the user experience of making sure the user realizes than only N-1 out
of N nodes were queried.
Thanks
Joe
> Unable to query provenance when node goes down
> ----------------------------------------------
>
> Key: NIFI-599
> URL: https://issues.apache.org/jira/browse/NIFI-599
> Project: Apache NiFi
> Issue Type: Bug
> Components: Core Framework
> Affects Versions: 0.0.2
> Environment: Configuration: multiple-node cluster
> Reporter: Miriam Kimball
> Priority: Minor
>
> If a node dies (ungracefully) from the cluster, users are unable to query
> provenance for other (still running) nodes.
> Steps to repeat:
> 1. Set up clustered config with at least 2 modes.
> 2. In web UI, do a Provenance query.
> 3. Take down one node ungracefully (kill it).
> 4. Do a Provenance query again.
> ---> Notice that an error is displayed: Node localhost:8081 is unable to
> fulfill this request due to: Unexpected Response Code 500.
> The user should be able to query the remaining nodes, but is prevented from
> doing so.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
