Check your configs on nifi2. I don't believe that NiFi is starting two
instances of Zookeeper but the ports configured are unintentionally configured
to overlap ie. Ports used twice in different configs where they should be
different.
It may be that your zookeeper.properties has:
clientPort=2180
...
server.1=nifi1.com:2180:3888
server.2=nifi2.com:2180:3888
where it should be:
clientPort=2180
...
server.1=nifi1.com:2888:3888
server.2=nifi2.com:2888:3888
noticing that the server.1 and server.2 ranges don't overlap with the client
port.
Not sure if this helps, but the following is the relevant config that I have
for my NiFi cluster nodes that run on the SAME machine where nifi1.com and
nifi2.com are configured in /etc/hosts:
nifi1/conf
zookeeper.properties
- clientPort=2180
- server.1=nifi1.com:2888:3888
- server.2=nifi2.com:2888:3888
nifi.properties
- nifi.remote.input.host=nifi1.com
- nifi.remote.input.socket.port=10440
- nifi.web.http.host=nifi1.com
- nifi.web.http.port=9550
- nifi.cluster.node.address=nifi1.com
- nifi.cluster.node.protocol.port=11440
nifi1/state/zookeeper
/myid (file contents = "1")
/state-management.xml (no changes required)
/version-2/
nifi2/conf
zookeeper.properties
- clientPort=2181
- server.1=nifi1.com:2888:3888
- server.2=nifi2.com:2888:3888
nifi.properties
- nifi.remote.input.host=nifi2.com
- nifi.remote.input.socket.port=10441
- nifi.web.http.host=nifi2.com
- nifi.web.http.port=9551
- nifi.cluster.node.address=nifi2.com
- nifi.cluster.node.protocol.port=11441
nifi2/state/zookeeper
/myid (file contents = "2")
/state-management.xml (no changes required)
/version-2/
Nathan
On 10/2/18, 2:07 AM, "Phil H" <gippyp...@gmail.com> wrote:
Hi Andy,
Thanks for the additional info. I think I saw a link to that while
searching but was wary since it was such an old version.
I have two VMs (nifi1, and nifi2) both running NiFi with identical configs,
and trying to use the inbuilt ZK to cluster them.
If I only mention a single machine within the config (eg: if nifi1 doesn’t
refer to nifi2, or visa versa) I don’t get any start up errors.
Phil
From: Andy LoPresto
Sent: Tuesday, 2 October 2018 1:00 PM
To: dev@nifi.apache.org
Subject: Re: Zookeeper - help!
Hi Phil,
Nathan’s advice is correct but I think he was assuming all other
configurations are correct as well. Are you trying to run both NiFi nodes and
ZK instances on the same machine? In that case you will have to ensure that the
ports in use are different for each service so they don’t conflict. Setting
them all to the same value only works if each service is running on an
independent physical machine, virtual machine, or container.
I find Pierre’s guide [1] to be a helpful step-by-step instruction list as
well as a good explanation of how the clustering concepts work in practice.
When you get that working, and you’re ready to set up a secure cluster, he has
a follow-on guide for that as well [2]. Even as someone who has set up many
clustered instances of NiFi, I use his guides regularly to ensure I haven’t
forgotten a step.
They were originally written for versions 1.0.0 and 1.1.0, but the only
thing that has changed is the authorizer configuration for the secure instances
(you’ll need to put the Initial Admin Identity and Node Identities in two
locations in the authorizers.xml file instead of just once).
Hopefully this helps you get a working cluster up and running so you can
experiment. Good luck.
[1] https://pierrevillard.com/2016/08/13/apache-nifi-1-0-0-cluster-setup/
[2]
https://pierrevillard.com/2016/11/29/apache-nifi-1-1-0-secured-cluster-setup/
Andy LoPresto
alopre...@apache.org
alopresto.apa...@gmail.com
PGP Fingerprint: 70EC B3E5 98A6 5A3F D3C4 BACE 3C6E F65B 2F7D EF69
On Oct 1, 2018, at 2:45 PM, Phil H <gippyp...@gmail.com> wrote:
Thanks Nathan,
I changed the protocol.port to 10002 on both servers.
On server 1, I now just see endless copies of the second error from my
original message (“KeeperException$ConnectionLossException: KeeperErrorCode =
ConnectionLoss”) – I don’t know if that’s normal when there’s only a single
member of a cluster alive and running? Seems like the logs will fill up very
quickly if it is!
On server 2, I get a bind exception on the Zookeeper client port. It
doesn’t matter what I set it to (In this example, I changed it to 10500) I
always get the same result. If I run netstat when nifi isn’t running, there’s
nothing listening on the port. It’s like NiFi is starting two Zookeeper
instances?! There’s no repeat of this in the start up sequence though. Both
servers are running completely vanilla 1.6.0 – I don’t even have any flow
defined yet, this is purely for teaching myself clustering config – so I don’t
know why one is behaving differently to the other.
2018-10-02 17:36:31,610 INFO [QuorumPeer[myid=2]/0.0.0.0:10500]
o.a.zookeeper.server.ZooKeeperServer Created server with tickTime 2000
minSessionTimeout 4000 maxSessionTimeout 40000 datadir
./state/zookeeper/version-2 snapdir ./state/zookeeper/version-2
2018-10-02 17:36:31,612 ERROR [QuorumPeer[myid=2]/0.0.0.0:10500]
o.apache.zookeeper.server.quorum.Leader Couldn't bind to
nifi2.domain/192.168.10.102:10500
java.net.BindException: Address already in use (Bind failed)
at java.net.PlainSocketImpl.socketBind(Native Method)
at
java.net.AbstractPlainSocketImpl.bind(AbstractPlainSocketImpl.java:387)
at java.net.ServerSocket.bind(ServerSocket.java:375)
at java.net.ServerSocket.bind(ServerSocket.java:329)
at org.apache.zookeeper.server.quorum.Leader.<init>(Leader.java:193)
at
org.apache.zookeeper.server.quorum.QuorumPeer.makeLeader(QuorumPeer.java:605)
at
org.apache.zookeeper.server.quorum.QuorumPeer.run(QuorumPeer.java:798)
From: Nathan Gough
Sent: Tuesday, 2 October 2018 2:22 AM
To: dev@nifi.apache.org
Subject: Re: Zookeeper - help!
Hi Phil,
One thing I notice with your config is that the cluster.node.protol.port
and the zookeeper ports are the same - these should not be the same.
Node.protocol.port is used by NiFi cluster to communicate between nodes, the
zookeeper.connect.string port should be the port that zookeeper service is
listening on. The zookeeper port is configured by the clientPort property in
the zookeeper.properties file. This would make your connect string:
'nifi.zookeeper.connect.string=nifi1.domain:2180,nifi2.domain:2180', where 2180
is whatever clientPort is configured.
You can read more about how NiFi uses Zookeeper and how to configure it
here:
https://nifi.apache.org/docs/nifi-docs/html/administration-guide.html#state_management.
Let us know what happens once these properties are configured correctly.
Nathan
On 9/30/18, 11:07 PM, "Phil H" <gippyp...@gmail.com> wrote:
Hi guys,
Pulling my hair out trying to solve my Zookeeper problems. I have two
1.6.0 servers that I am trying to cluster.
Here is the except from the properties files – all other properties are
default so omitted for clarity. The servers are set up to run HTTPS, and the
interface works via the browser, so I believe the certificates are correctly
installed.
Server nifi1.domain:
nifi.cluster.is.node=true
nifi.cluster.node.address=nifi1.domain
nifi.cluster.node.protocol.port=10000
nifi.zookeeper.connect.string=nifi2.domain:10000,nifi1.domain:10000
nifi.zookeeper.root.node=/nifi
Server nifi2.domain:
nifi.cluster.is.node=true
nifi.cluster.node.address=nifi2.domain
nifi.cluster.node.protocol.port=10000
nifi.zookeeper.connect.string=nifi1.domain:10000,nifi2.domain:10000
nifi.zookeeper.root.node=/nifi
I am getting these errors (this is from server 2, but seeing the same on
server 1 apart from a different address, of course):
2018-10-01 20:54:16,332 INFO [main]
org.apache.nifi.io.socket.SocketListener Now listening for connections from
nodes on port 10000
2018-10-01 20:54:16,381 INFO [main]
o.apache.nifi.controller.FlowController Successfully synchronized controller
with proposed flow
2018-10-01 20:54:16,435 INFO [main]
o.a.nifi.controller.StandardFlowService Connecting Node: nifi2.domain:443
2018-10-01 20:54:16,769 ERROR [Process Cluster Protocol Request-1]
o.a.nifi.security.util.CertificateUtils The incoming request did not contain
client certificates and thus the DN cannot be extracted. Check that the other
endpoint is providing a complete client certificate chain
2018-10-01 20:54:16,771 WARN [Process Cluster Protocol Request-1]
o.a.n.c.p.impl.SocketProtocolListener Failed processing protocol message from
nifi2 due to org.apache.nifi.cluster.protocol.ProtocolException:
java.security.cert.CertificateException:
javax.net.ssl.SSLPeerUnverifiedException: peer not authenticated
org.apache.nifi.cluster.protocol.ProtocolException:
java.security.cert.CertificateException:
javax.net.ssl.SSLPeerUnverifiedException: peer not authenticated
at
org.apache.nifi.cluster.protocol.impl.SocketProtocolListener.getRequestorDN(SocketProtocolListener.java:225)
at
org.apache.nifi.cluster.protocol.impl.SocketProtocolListener.dispatchRequest(SocketProtocolListener.java:131)
at
org.apache.nifi.io.socket.SocketListener$2$1.run(SocketListener.java:136)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: java.security.cert.CertificateException:
javax.net.ssl.SSLPeerUnverifiedException: peer not authenticated
at
org.apache.nifi.security.util.CertificateUtils.extractPeerDNFromClientSSLSocket(CertificateUtils.java:314)
at
org.apache.nifi.security.util.CertificateUtils.extractPeerDNFromSSLSocket(CertificateUtils.java:269)
at
org.apache.nifi.cluster.protocol.impl.SocketProtocolListener.getRequestorDN(SocketProtocolListener.java:223)
... 5 common frames omitted
Caused by: javax.net.ssl.SSLPeerUnverifiedException: peer not
authenticated
at
sun.security.ssl.SSLSessionImpl.getPeerCertificates(SSLSessionImpl.java:440)
at
org.apache.nifi.security.util.CertificateUtils.extractPeerDNFromClientSSLSocket(CertificateUtils.java:299)
... 7 common frames omitted
2018-10-01 20:54:32,249 INFO [Curator-Framework-0]
o.a.c.f.state.ConnectionStateManager State change: SUSPENDED
2018-10-01 20:54:32,250 ERROR [Curator-Framework-0]
o.a.c.f.imps.CuratorFrameworkImpl Background operation retry gave up
org.apache.zookeeper.KeeperException$ConnectionLossException:
KeeperErrorCode = ConnectionLoss
at
org.apache.zookeeper.KeeperException.create(KeeperException.java:99)
at
org.apache.curator.framework.imps.CuratorFrameworkImpl.checkBackgroundRetry(CuratorFrameworkImpl.java:728)
at
org.apache.curator.framework.imps.CuratorFrameworkImpl.performBackgroundOperation(CuratorFrameworkImpl.java:857)
at
org.apache.curator.framework.imps.CuratorFrameworkImpl.backgroundOperationsLoop(CuratorFrameworkImpl.java:809)
at
org.apache.curator.framework.imps.CuratorFrameworkImpl.access$300(CuratorFrameworkImpl.java:64)
at
org.apache.curator.framework.imps.CuratorFrameworkImpl$4.call(CuratorFrameworkImpl.java:267)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180)
at
java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
