nifi-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jon Logan <jmlo...@buffalo.edu>
Subject Re: Cluster Crashing When Nodes Unresponsive?
Date Wed, 02 May 2018 16:22:53 GMT
We're on 1.4.0. I suspect it's because our cluster timeouts were too
high...we know some nodes were dying, but it seemed bizarre that it would
take down the whole cluster. ex. we restarted one problematic node and the
entire cluster came back to life. Should that unresponsive node failure not
be contained to just that node?

On Tue, May 1, 2018 at 8:29 PM, Juan Sequeiros <hellojuan@gmail.com> wrote:

> Hi what version of NIFI are you on?
> In my experience it’s usually resource issue and most of the time it’s
> disk I/O.
>
>
> On Tue, May 1, 2018 at 4:43 PM Jon Logan <jmlogan@buffalo.edu> wrote:
>
>> We are running into issues where if a node is unresponsive for whatever
>> reason, the entire cluster seems to grind to a halt. I suspect that this is
>> because it can't replicate properly due to the node crashing, etc...is this
>> known/expected behavior? We are connecting a Nifi cluster back to itself
>> via a RPG to distribute the load, which I think is a standard design
>> practice?
>>
>> I figured that node would be penalized and/or disconnected, and life
>> would move on for the rest of the cluster?
>>
>>
>> Thanks!
>>
>

Mime
View raw message