nifi-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Matthew Clarke <matt.clarke....@gmail.com>
Subject Re: Nifi cluster features - Questions
Date Mon, 11 Jan 2016 22:05:52 GMT
Chakri,
     Your original option was to use a RPG (Site-to-Site) to redistribute
data received on the primary Node only to every node in your cluster in a
load-balanced fashion. In this setup every connected Node will get a
portion of the data and there is no way to specify which node gets what
particular piece of data.  This dataflwo would look like this:
[image: Inline image 1]
    Data is received by the listenHTTP processor on the primary Node only.
It is then sent to a RPG that was added using the URL of the NCM for your
10 Node cluster (the cluster also running the before mentioned primary node
listenHTTP processor). There will also be an input port which runs on every
Node.  each Node will receive a load-balanced distribution of the data sent
to that RPG.

Hope these illustration help.

Thanks,
Matt

On Mon, Jan 11, 2016 at 4:50 PM, Matthew Clarke <matt.clarke.138@gmail.com>
wrote:

> Chakri,
>         All data is received on the primary Node only via the initial
> listenHTTP.  Some routing tales place to send some data to a particular 5
> nodes and other data to the other 5 nodes.  The postHTTP processor are
> configured to send to a specific Node in your cluster using the same target
> port number. A single ListenHTTP processor lives then runs on every Node
> configured to use that target port number.
>
> Thanks,
> Matt
>
> On Mon, Jan 11, 2016 at 4:47 PM, Matthew Clarke <matt.clarke.138@gmail.com
> > wrote:
>
>> Chakri,
>>             What Mark is saying is NiFI Remote Process Group (RPG) also
>> known as Site-to-Site will load-balance delivery data to all nodes in a
>> cluster.  It can not be configured to balance data to only a subset of a
>> nodes in a cluster.  If this is the strategy you want to deploy, a
>> different approach must be taken (one that does not use Site-to-Site).
>> Here is a NiFI diagram of one such approach using your example of a 10 node
>> cluster:
>>
>> [image: Inline image 1]
>>
>>
>>
>> On Mon, Jan 11, 2016 at 4:16 PM, Chakrader Dewaragatla <
>> Chakrader.Dewaragatla@lifelock.com> wrote:
>>
>>> Mark - Correct me if I understood right.
>>>
>>> Curl post from some application —> Configure Listen http (on primary
>>> node) --> Post http with Data flow file (On primary node?)  --> Post to
>>> site-to-site end point —> This intern distribute load to both slaves.
>>>
>>> Thanks,
>>> -Chakri
>>>
>>> From: Mark Payne <markap14@hotmail.com>
>>> Reply-To: "users@nifi.apache.org" <users@nifi.apache.org>
>>> Date: Monday, January 11, 2016 at 12:29 PM
>>>
>>> To: "users@nifi.apache.org" <users@nifi.apache.org>
>>> Subject: Re: Nifi cluster features - Questions
>>>
>>> Chakri,
>>>
>>> At this time, your only options are to run the processors on all nodes
>>> or a single node (Primary Node). There's no way to really group nodes
>>> together and say "only run on this set of nodes."
>>>
>>> One option is to have a ListenHTTP Processor and then push data to that
>>> NiFi via PostHTTP (configure it to send FlowFile attributes along). By
>>> doing this, you could set up the sending NiFi
>>> to only deliver data to two nodes. You could then have a different set
>>> of data going to a different two nodes, etc. by the way that you configure
>>> which data goes to which PostHTTP Processor.
>>>
>>> Does this give you what you need?
>>>
>>>
>>> On Jan 11, 2016, at 3:20 PM, Chakrader Dewaragatla <
>>> Chakrader.Dewaragatla@lifelock.com> wrote:
>>>
>>> Thanks Mark. I will look into it.
>>>
>>> Couple of questions:
>>>
>>>
>>>    - Going back to my earlier question, In a nifi cluster with two
>>>    slaves and NCM how do I make two slaves accept and process the incoming
>>>    flowfile in distibuted fashion. Site to site is the only way to go ?
>>>    In our use case, we have http listener running on primary node and
>>>    putfile processor should run on two slaves in distributed fashion.
>>>
>>>    It is more like a new (or existing) feature.
>>>     - In a nifi cluster setup, can we group the machines and set
>>>    site-to-site to individual group.
>>>     For instance I have 10 node cluster, can I group them into 5 groups
>>>    with two nodes each. Run processors on dedicated group (using site to site
>>>    or other means).
>>>
>>> Thanks,
>>> -Chakri
>>>
>>> From: Mark Payne <markap14@hotmail.com>
>>> Reply-To: "users@nifi.apache.org" <users@nifi.apache.org>
>>> Date: Monday, January 11, 2016 at 5:24 AM
>>> To: "users@nifi.apache.org" <users@nifi.apache.org>
>>> Subject: Re: Nifi cluster features - Questions
>>>
>>> Chakri,
>>>
>>> This line in the logs is particularly interesting (on primary node):
>>>
>>> 2016-01-10 22:52:39,459 INFO [Timer-Driven Process Thread-7]
>>> o.a.n.r.c.socket.EndpointConnectionPool New Weighted Distribution of Nodes:
>>> Node[i-c894e249.dev.aws.lifelock.ad:0] will receive 100.0% of data
>>>
>>>
>>> This indicates that all of the site-to-site data will go to the host
>>> i-c894e249.dev.aws.lifelock.ad. Moreover, because that is the only node
>>> listed, this means
>>> that the NCM responded, indicating that this is the only node in the
>>> cluster that is currently connected and has site-to-site enabled. Can you
>>> double-check the nifi.properties
>>> file on the Primary Node and verify that the "
>>> nifi.remote.input.socket.port" is property is specified, and that the "
>>> nifi.remote.input.secure" property is set to "false"?
>>> Of note is that if the "nifi.remote.input.secure" property is set to
>>> true, but keystore and truststore are not specified, then site-to-site will
>>> be disabled (there would be a warning
>>> in the log in this case).
>>>
>>> If you can verify that both of those properties are set properly on both
>>> nodes, then we can delve in further, but probably best to start by
>>> double-checking the easy things :)
>>>
>>> Thanks
>>> -Mark
>>>
>>>
>>> On Jan 10, 2016, at 5:55 PM, Chakrader Dewaragatla <
>>> Chakrader.Dewaragatla@lifelock.com> wrote:
>>>
>>> Bryan – Here are the logs :
>>> I have 5 sec flow file.
>>>
>>> On primary node (No data coming in)
>>>
>>> 2016-01-10 22:52:36,322 INFO [Clustering Tasks Thread-1]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:36,146 and sent at 2016-01-10 22:52:36,322; send took 0 millis
>>> 2016-01-10 22:52:36,476 INFO [Flow Service Tasks Thread-2]
>>> o.a.nifi.controller.StandardFlowService Saved flow controller
>>> org.apache.nifi.controller.FlowController@5dff8cbf // Another save
>>> pending = false
>>> 2016-01-10 22:52:39,450 INFO [pool-26-thread-16]
>>> o.a.n.c.s.TimerDrivenSchedulingAgent Scheduled
>>> GenerateFlowFile[id=6efbcd69-0b82-4ea2-a90d-01b39efaf3db] to run with 1
>>> threads
>>> 2016-01-10 22:52:39,459 INFO [Timer-Driven Process Thread-7]
>>> o.a.n.r.c.socket.EndpointConnectionPool New Weighted Distribution of Nodes:
>>> Node[i-c894e249.dev.aws.lifelock.ad:0] will receive 100.0% of data
>>> 2016-01-10 22:52:39,480 INFO [Flow Service Tasks Thread-2]
>>> o.a.nifi.controller.StandardFlowService Saved flow controller
>>> org.apache.nifi.controller.FlowController@5dff8cbf // Another save
>>> pending = false
>>> 2016-01-10 22:52:39,576 INFO [Clustering Tasks Thread-2]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:39,452 and sent at 2016-01-10 22:52:39,576; send took 1 millis
>>> 2016-01-10 22:52:39,662 INFO [Timer-Driven Process Thread-7]
>>> o.a.nifi.remote.StandardRemoteGroupPort
>>> RemoteGroupPort[name=nifi-input,target=http://10.228.68.73:8080/nifi]
>>> Successfully sent
>>> [StandardFlowFileRecord[uuid=f6ff266d-e03f-4a8e-af5a-1455dd433ff4,claim=StandardContentClaim
>>> [resourceClaim=StandardResourceClaim[id=1452456659209-1, container=default,
>>> section=1], offset=1980, length=20],offset=0,name=275238507698589,size=20]]
>>> (20 bytes) to nifi://i-c894e249.dev.aws.lifelock.ad:10880 in 50
>>> milliseconds at a rate of 392 bytes/sec
>>> 2016-01-10 22:52:41,327 INFO [Clustering Tasks Thread-1]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:41,147 and sent at 2016-01-10 22:52:41,327; send took 0 millis
>>> 2016-01-10 22:52:44,524 INFO [Timer-Driven Process Thread-1]
>>> o.a.nifi.remote.StandardRemoteGroupPort
>>> RemoteGroupPort[name=nifi-input,target=http://10.228.68.73:8080/nifi]
>>> Successfully sent
>>> [StandardFlowFileRecord[uuid=effbc026-98d2-4548-9069-f95d57c8bf4b,claim=StandardContentClaim
>>> [resourceClaim=StandardResourceClaim[id=1452456659209-1, container=default,
>>> section=1], offset=2000, length=20],offset=0,name=275243509297560,size=20]]
>>> (20 bytes) to nifi://i-c894e249.dev.aws.lifelock.ad:10880 in 51
>>> milliseconds at a rate of 391 bytes/sec
>>> 2016-01-10 22:52:45,092 INFO [Process NCM Request-2]
>>> o.a.n.c.p.impl.SocketProtocolListener Received request
>>> 8ecc76f9-e978-4e9b-a8ed-41a47647d5bd from 10.228.68.73
>>> 2016-01-10 22:52:45,094 INFO [Process NCM Request-2]
>>> o.a.nifi.controller.StandardFlowService Received flow request message from
>>> manager.
>>> 2016-01-10 22:52:45,094 INFO [Process NCM Request-2]
>>> o.a.n.c.p.impl.SocketProtocolListener Finished processing request
>>> 8ecc76f9-e978-4e9b-a8ed-41a47647d5bd (type=FLOW_REQUEST, length=331 bytes)
>>> in 61 millis
>>> 2016-01-10 22:52:46,391 INFO [Clustering Tasks Thread-1]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:46,148 and sent at 2016-01-10 22:52:46,391; send took 60 millis
>>> 2016-01-10 22:52:48,470 INFO [Provenance Maintenance Thread-3]
>>> o.a.n.p.PersistentProvenanceRepository Created new Provenance Event Writers
>>> for events starting with ID 301
>>> 2016-01-10 22:52:48,580 INFO [Provenance Repository Rollover Thread-2]
>>> o.a.n.p.PersistentProvenanceRepository Successfully merged 16 journal files
>>> (6 records) into single Provenance Log File
>>> ./provenance_repository/295.prov in 111 milliseconds
>>> 2016-01-10 22:52:48,580 INFO [Provenance Repository Rollover Thread-2]
>>> o.a.n.p.PersistentProvenanceRepository Successfully Rolled over Provenance
>>> Event file containing 8 records
>>> 2016-01-10 22:52:49,517 INFO [Timer-Driven Process Thread-10]
>>> o.a.nifi.remote.StandardRemoteGroupPort
>>> RemoteGroupPort[name=nifi-input,target=http://10.228.68.73:8080/nifi]
>>> Successfully sent
>>> [StandardFlowFileRecord[uuid=505bef8e-15e6-4345-b909-cb3be21275bd,claim=StandardContentClaim
>>> [resourceClaim=StandardResourceClaim[id=1452456659209-1, container=default,
>>> section=1], offset=2020, length=20],offset=0,name=275248510432074,size=20]]
>>> (20 bytes) to nifi://i-c894e249.dev.aws.lifelock.ad:10880 in 50
>>> milliseconds at a rate of 392 bytes/sec
>>> 2016-01-10 22:52:51,395 INFO [Clustering Tasks Thread-3]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:51,150 and sent at 2016-01-10 22:52:51,395; send took 0 millis
>>> 2016-01-10 22:52:54,326 INFO [NiFi Web Server-22]
>>> o.a.n.c.s.TimerDrivenSchedulingAgent Stopped scheduling
>>> StandardRootGroupPort[name=nifi-input,id=392bfcc3-dfc2-4497-8148-8128336856fa]
>>> to run
>>> 2016-01-10 22:52:54,353 INFO [NiFi Web Server-26]
>>> o.a.n.c.s.TimerDrivenSchedulingAgent Stopped scheduling
>>> PutFile[id=2a2c47e1-a4cf-4c32-ba17-d195af3c2a1b] to run
>>> 2016-01-10 22:52:54,377 INFO [NiFi Web Server-25]
>>> o.a.n.c.s.TimerDrivenSchedulingAgent Stopped scheduling
>>> GenerateFlowFile[id=6efbcd69-0b82-4ea2-a90d-01b39efaf3db] to run
>>> 2016-01-10 22:52:54,397 INFO [Clustering Tasks Thread-2]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:54,379 and sent at 2016-01-10 22:52:54,397; send took 0 millis
>>> 2016-01-10 22:52:54,488 INFO [Flow Service Tasks Thread-2]
>>> o.a.nifi.controller.StandardFlowService Saved flow controller
>>> org.apache.nifi.controller.FlowController@5dff8cbf // Another save
>>> pending = false
>>> 2016-01-10 22:52:56,399 INFO [Clustering Tasks Thread-1]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:56,151 and sent at 2016-01-10 22:52:56,399; send took 0 millis
>>>
>>>
>>> On Secondary node (Data coming in)
>>>
>>> 2016-01-10 22:52:43,896 INFO [pool-18-thread-1]
>>> o.a.n.c.r.WriteAheadFlowFileRepository Successfully checkpointed FlowFile
>>> Repository with 0 records in 88 milliseconds
>>> 2016-01-10 22:52:44,524 INFO [Timer-Driven Process Thread-3]
>>> o.a.n.r.p.s.SocketFlowFileServerProtocol
>>> SocketFlowFileServerProtocol[CommsID=e3151c71-9c43-4179-a69d-bc1e1b94b573]
>>> Successfully received
>>> [StandardFlowFileRecord[uuid=614a656d-965b-4915-95f7-ee59e049ea20,claim=StandardContentClaim
>>> [resourceClaim=StandardResourceClaim[id=1452457702480-1, container=default,
>>> section=1], offset=1960, length=20],offset=0,name=275243509297560,size=20]]
>>> (20 bytes) from Peer[url=nifi://10.228.68.106:40611] in 51 milliseconds
>>> at a rate of 387 bytes/sec
>>> 2016-01-10 22:52:44,534 INFO [Timer-Driven Process Thread-1]
>>> o.a.nifi.processors.standard.PutFile
>>> PutFile[id=2a2c47e1-a4cf-4c32-ba17-d195af3c2a1b] Produced copy of
>>> StandardFlowFileRecord[uuid=614a656d-965b-4915-95f7-ee59e049ea20,claim=StandardContentClaim
>>> [resourceClaim=StandardResourceClaim[id=1452457702480-1, container=default,
>>> section=1], offset=1960, length=20],offset=0,name=275243509297560,size=20]
>>> at location /root/putt/275243509297560
>>> 2016-01-10 22:52:44,671 INFO [Provenance Maintenance Thread-3]
>>> o.a.n.p.PersistentProvenanceRepository Created new Provenance Event Writers
>>> for events starting with ID 17037
>>> 2016-01-10 22:52:44,727 INFO [Provenance Repository Rollover Thread-1]
>>> o.a.n.p.PersistentProvenanceRepository Successfully merged 16 journal files
>>> (6 records) into single Provenance Log File
>>> ./provenance_repository/17031.prov in 56 milliseconds
>>> 2016-01-10 22:52:44,727 INFO [Provenance Repository Rollover Thread-1]
>>> o.a.n.p.PersistentProvenanceRepository Successfully Rolled over Provenance
>>> Event file containing 10 records
>>> 2016-01-10 22:52:45,034 INFO [Process NCM Request-2]
>>> o.a.n.c.p.impl.SocketProtocolListener Received request
>>> e288a3eb-28fb-48cf-9f4b-bc36acb810bb from 10.228.68.73
>>> 2016-01-10 22:52:45,036 INFO [Process NCM Request-2]
>>> o.a.nifi.controller.StandardFlowService Received flow request message from
>>> manager.
>>> 2016-01-10 22:52:45,036 INFO [Process NCM Request-2]
>>> o.a.n.c.p.impl.SocketProtocolListener Finished processing request
>>> e288a3eb-28fb-48cf-9f4b-bc36acb810bb (type=FLOW_REQUEST, length=331 bytes)
>>> in 76 millis
>>> 2016-01-10 22:52:45,498 INFO [Clustering Tasks Thread-2]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:45,421 and sent at 2016-01-10 22:52:45,498; send took 0 millis
>>> 2016-01-10 22:52:49,518 INFO [Timer-Driven Process Thread-6]
>>> o.a.n.r.p.s.SocketFlowFileServerProtocol
>>> SocketFlowFileServerProtocol[CommsID=e3151c71-9c43-4179-a69d-bc1e1b94b573]
>>> Successfully received
>>> [StandardFlowFileRecord[uuid=a6986405-1f15-4233-a06f-1b9ce50c0e24,claim=StandardContentClaim
>>> [resourceClaim=StandardResourceClaim[id=1452457702480-1, container=default,
>>> section=1], offset=1980, length=20],offset=0,name=275248510432074,size=20]]
>>> (20 bytes) from Peer[url=nifi://10.228.68.106:40611] in 51 milliseconds
>>> at a rate of 387 bytes/sec
>>> 2016-01-10 22:52:49,520 INFO [Timer-Driven Process Thread-8]
>>> o.a.nifi.processors.standard.PutFile
>>> PutFile[id=2a2c47e1-a4cf-4c32-ba17-d195af3c2a1b] Produced copy of
>>> StandardFlowFileRecord[uuid=a6986405-1f15-4233-a06f-1b9ce50c0e24,claim=StandardContentClaim
>>> [resourceClaim=StandardResourceClaim[id=1452457702480-1, container=default,
>>> section=1], offset=1980, length=20],offset=0,name=275248510432074,size=20]
>>> at location /root/putt/275248510432074
>>> 2016-01-10 22:52:50,561 INFO [Clustering Tasks Thread-1]
>>> org.apache.nifi.cluster.heartbeat Heartbeat created at 2016-01-10
>>> 22:52:50,423 and sent at 2016-01-10 22:52:50,561; send took 59 millis
>>> From: Bryan Bende <bbende@gmail.com>
>>> Reply-To: "users@nifi.apache.org" <users@nifi.apache.org>
>>> Date: Sunday, January 10, 2016 at 2:43 PM
>>> To: "users@nifi.apache.org" <users@nifi.apache.org>
>>> Subject: Re: Nifi cluster features - Questions
>>>
>>> Chakri,
>>>
>>> Glad you got site-to-site working.
>>>
>>> Regarding the data distribution, I'm not sure why it is behaving that
>>> way. I just did a similar test running ncm, node1, and node2 all on my
>>> local machine, with GenerateFlowFile running every 10 seconds, and Input
>>> Port going to a LogAttribute, and I see it alternating between node1 and
>>> node2 logs every 10 seconds.
>>>
>>> Is there anything in your primary node logs
>>> (primary_node/logs/nifi-app.log) when you see the data on the other node?
>>>
>>> -Bryan
>>>
>>>
>>> On Sun, Jan 10, 2016 at 3:44 PM, Joe Witt <joe.witt@gmail.com> wrote:
>>>
>>>> Chakri,
>>>>
>>>> Would love to hear what you've learned and how that differed from the
>>>> docs themselves.  Site-to-site has proven difficult to setup so we're
>>>> clearly not there yet in having the right operator/admin experience.
>>>>
>>>> Thanks
>>>> Joe
>>>>
>>>> On Sun, Jan 10, 2016 at 3:41 PM, Chakrader Dewaragatla
>>>> <Chakrader.Dewaragatla@lifelock.com> wrote:
>>>> > I was able to get site-to-site work.
>>>> > I tried to follow your instructions to send data distribute across the
>>>> > nodes.
>>>> >
>>>> > GenerateFlowFile (On Primary) —> RPG
>>>> > RPG —> Input Port   —> Putfile (Time driven scheduling)
>>>> >
>>>> > However, data is only written to one slave (Secondary slave). Primary
>>>> slave
>>>> > has not data.
>>>> >
>>>> > Image screenshot :
>>>> > http://tinyurl.com/jjvjtmq
>>>> >
>>>> > From: Chakrader Dewaragatla <chakrader.dewaragatla@lifelock.com>
>>>> > Date: Sunday, January 10, 2016 at 11:26 AM
>>>> >
>>>> > To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> > Subject: Re: Nifi cluster features - Questions
>>>> >
>>>> > Bryan – Thanks – I am trying to setup site-to-site.
>>>> > I have two slaves and one NCM.
>>>> >
>>>> > My properties as follows :
>>>> >
>>>> > On both Slaves:
>>>> >
>>>> > nifi.remote.input.socket.port=10880
>>>> > nifi.remote.input.secure=false
>>>> >
>>>> > On NCM:
>>>> > nifi.remote.input.socket.port=10880
>>>> > nifi.remote.input.secure=false
>>>> >
>>>> > When I try drop remote process group (with http://<NCM
>>>> IP>:8080/nifi), I see
>>>> > error as follows for two nodes.
>>>> >
>>>> > [<Slave1 ip>:8080] - Remote instance is not allowed for Site to
Site
>>>> > communication
>>>> > [<Slave2 ip>:8080] - Remote instance is not allowed for Site to
Site
>>>> > communication
>>>> >
>>>> > Do you have insight why its trying to connecting 8080 on slaves ?
>>>> When do
>>>> > 10880 port come into the picture ? I remember try setting site to
>>>> site few
>>>> > months back and succeeded.
>>>> >
>>>> > Thanks,
>>>> > -Chakri
>>>> >
>>>> >
>>>> >
>>>> > From: Bryan Bende <bbende@gmail.com>
>>>> > Reply-To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> > Date: Saturday, January 9, 2016 at 11:22 AM
>>>> > To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> > Subject: Re: Nifi cluster features - Questions
>>>> >
>>>> > The sending node (where the remote process group is) will distribute
>>>> the
>>>> > data evenly across the two nodes, so an individual file will only be
>>>> sent to
>>>> > one of the nodes. You could think of it as if a separate NiFi
>>>> instance was
>>>> > sending directly to a two node cluster, it would be evenly
>>>> distributing the
>>>> > data across the two nodes. In this case it just so happens to all be
>>>> with in
>>>> > the same cluster.
>>>> >
>>>> > The most common use case for this scenario is the List and Fetch
>>>> processors
>>>> > like HDFS. You can perform the listing on primary node, and then
>>>> distribute
>>>> > the results so the fetching takes place on all nodes.
>>>> >
>>>> > On Saturday, January 9, 2016, Chakrader Dewaragatla
>>>> > <Chakrader.Dewaragatla@lifelock.com> wrote:
>>>> >>
>>>> >> Bryan – Thanks, how do the nodes distribute the load for a input
>>>> port. As
>>>> >> port is open and listening on two nodes,  does it copy same files
on
>>>> both
>>>> >> the nodes?
>>>> >> I need to try this setup to see the results, appreciate your help.
>>>> >>
>>>> >> Thanks,
>>>> >> -Chakri
>>>> >>
>>>> >> From: Bryan Bende <bbende@gmail.com>
>>>> >> Reply-To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> >> Date: Friday, January 8, 2016 at 3:44 PM
>>>> >> To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> >> Subject: Re: Nifi cluster features - Questions
>>>> >>
>>>> >> Hi Chakri,
>>>> >>
>>>> >> I believe the DistributeLoad processor is more for load balancing
>>>> when
>>>> >> sending to downstream systems. For example, if you had two HTTP
>>>> endpoints,
>>>> >> you could have the first relationship from DistributeLoad going
to a
>>>> >> PostHTTP that posts to endpoint #1, and the second relationship
>>>> going to a
>>>> >> second PostHTTP that goes to endpoint #2.
>>>> >>
>>>> >> If you want to distribute the data with in the cluster, then you
>>>> need to
>>>> >> use site-to-site. The way you do this is the following...
>>>> >>
>>>> >> - Add an Input Port connected to your PutFile.
>>>> >> - Add GenerateFlowFile scheduled on primary node only, connected
to a
>>>> >> Remote Process Group. The Remote Process Group should be connected
>>>> to the
>>>> >> Input Port from the previous step.
>>>> >>
>>>> >> So both nodes have an input port listening for data, but only the
>>>> primary
>>>> >> node produces a FlowFile and sends it to the RPG which then
>>>> re-distributes
>>>> >> it back to one of the Input Ports.
>>>> >>
>>>> >> In order for this to work you need to set
>>>> nifi.remote.input.socket.port in
>>>> >> nifi.properties to some available port, and you probably want
>>>> >> nifi.remote.input.secure=false for testing.
>>>> >>
>>>> >> -Bryan
>>>> >>
>>>> >>
>>>> >> On Fri, Jan 8, 2016 at 6:27 PM, Chakrader Dewaragatla
>>>> >> <Chakrader.Dewaragatla@lifelock.com> wrote:
>>>> >>>
>>>> >>> Mark – I have setup a two node cluster and tried the following
.
>>>> >>>  GenrateFlowfile processor (Run only on primary node) —>
>>>> DistributionLoad
>>>> >>> processor (RoundRobin)   —> PutFile
>>>> >>>
>>>> >>> >> The GetFile/PutFile will run on all nodes (unless you
schedule
>>>> it to
>>>> >>> >> run on primary node only).
>>>> >>> From your above comment, It should put file on two nodes. It
put
>>>> files on
>>>> >>> primary node only. Any thoughts ?
>>>> >>>
>>>> >>> Thanks,
>>>> >>> -Chakri
>>>> >>>
>>>> >>> From: Mark Payne <markap14@hotmail.com>
>>>> >>> Reply-To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> >>> Date: Wednesday, October 7, 2015 at 11:28 AM
>>>> >>>
>>>> >>> To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> >>> Subject: Re: Nifi cluster features - Questions
>>>> >>>
>>>> >>> Chakri,
>>>> >>>
>>>> >>> Correct - when NiFi instances are clustered, they do not transfer
>>>> data
>>>> >>> between the nodes. This is very different
>>>> >>> than you might expect from something like Storm or Spark, as
the key
>>>> >>> goals and design are quite different.
>>>> >>> We have discussed providing the ability to allow the user to
>>>> indicate
>>>> >>> that they want to have the framework
>>>> >>> do load balancing for specific connections in the background,
but
>>>> it's
>>>> >>> still in more of a discussion phase.
>>>> >>>
>>>> >>> Site-to-Site is simply the capability that we have developed
to
>>>> transfer
>>>> >>> data between one instance of
>>>> >>> NiFi and another instance of NiFi. So currently, if we want
to do
>>>> load
>>>> >>> balancing across the cluster, we would
>>>> >>> create a site-to-site connection (by dragging a Remote Process
>>>> Group onto
>>>> >>> the graph) and give that
>>>> >>> site-to-site connection the URL of our cluster. That way, you
can
>>>> push
>>>> >>> data to your own cluster, effectively
>>>> >>> providing a load balancing capability.
>>>> >>>
>>>> >>> If you were to just run ListenHTTP without setting it to Primary
>>>> Node,
>>>> >>> then every node in the cluster will be listening
>>>> >>> for incoming HTTP connections. So you could then use a simple
load
>>>> >>> balancer in front of NiFi to distribute the load
>>>> >>> across your cluster.
>>>> >>>
>>>> >>> Does this help? If you have any more questions we're happy to
help!
>>>> >>>
>>>> >>> Thanks
>>>> >>> -Mark
>>>> >>>
>>>> >>>
>>>> >>> On Oct 7, 2015, at 2:32 PM, Chakrader Dewaragatla
>>>> >>> <Chakrader.Dewaragatla@lifelock.com> wrote:
>>>> >>>
>>>> >>> Mark - Thanks for the notes.
>>>> >>>
>>>> >>> >> The other option would be to have a ListenHTTP processor
run on
>>>> >>> >> Primary Node only and then use Site-to-Site to distribute
the
>>>> data to other
>>>> >>> >> nodes.
>>>> >>> Lets say I have 5 node cluster and ListenHTTP processor on Primary
>>>> node,
>>>> >>> collected data on primary node is not transfered to other nodes
by
>>>> default
>>>> >>> for processing despite all nodes are part of one cluster?
>>>> >>> If ListenHTTP processor is running  as a dafult (with out explicit
>>>> >>> setting to run on primary node), how does the data transferred
to
>>>> rest of
>>>> >>> the nodes? Does site-to-site come in play when I make one processor
>>>> to run
>>>> >>> on primary node ?
>>>> >>>
>>>> >>> Thanks,
>>>> >>> -Chakri
>>>> >>>
>>>> >>> From: Mark Payne <markap14@hotmail.com>
>>>> >>> Reply-To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> >>> Date: Wednesday, October 7, 2015 at 7:00 AM
>>>> >>> To: "users@nifi.apache.org" <users@nifi.apache.org>
>>>> >>> Subject: Re: Nifi cluster features - Questions
>>>> >>>
>>>> >>> Hello Chakro,
>>>> >>>
>>>> >>> When you create a cluster of NiFi instances, each node in the
>>>> cluster is
>>>> >>> acting independently and in exactly
>>>> >>> the same way. I.e., if you have 5 nodes, all 5 nodes will run
>>>> exactly the
>>>> >>> same flow. However, they will be
>>>> >>> pulling in different data and therefore operating on different
data.
>>>> >>>
>>>> >>> So if you pull in 10 1-gig files from S3, each of those files
will
>>>> be
>>>> >>> processed on the node that pulled the data
>>>> >>> in. NiFi does not currently shuffle data around between nodes
in the
>>>> >>> cluster (you can use site-to-site to do
>>>> >>> this if you want to, but it won't happen automatically). If
you set
>>>> the
>>>> >>> number of Concurrent Tasks to 5, then
>>>> >>> you will have up to 5 threads running for that processor on
each
>>>> node.
>>>> >>>
>>>> >>> The only exception to this is the Primary Node. You can schedule
a
>>>> >>> Processor to run only on the Primary Node
>>>> >>> by right-clicking on the Processor, and going to the Configure
>>>> menu. In
>>>> >>> the Scheduling tab, you can change
>>>> >>> the Scheduling Strategy to Primary Node Only. In this case,
that
>>>> >>> Processor will only be triggered to run on
>>>> >>> whichever node is elected the Primary Node (this can be changed
in
>>>> the
>>>> >>> Cluster management screen by clicking
>>>> >>> the appropriate icon in the top-right corner of the UI).
>>>> >>>
>>>> >>> The GetFile/PutFile will run on all nodes (unless you schedule
it
>>>> to run
>>>> >>> on primary node only).
>>>> >>>
>>>> >>> If you are attempting to have a single input running HTTP and
then
>>>> push
>>>> >>> that out across the entire cluster to
>>>> >>> process the data, you would have a few options. First, you could
>>>> just use
>>>> >>> an HTTP Load Balancer in front of NiFi.
>>>> >>> The other option would be to have a ListenHTTP processor run
on
>>>> Primary
>>>> >>> Node only and then use Site-to-Site
>>>> >>> to distribute the data to other nodes.
>>>> >>>
>>>> >>> For more info on site-to-site, you can see the Site-to-Site
section
>>>> of
>>>> >>> the User Guide at
>>>> >>>
>>>> http://nifi.apache.org/docs/nifi-docs/html/user-guide.html#site-to-site
>>>> >>>
>>>> >>> If you have any more questions, let us know!
>>>> >>>
>>>> >>> Thanks
>>>> >>> -Mark
>>>> >>>
>>>> >>> On Oct 7, 2015, at 2:33 AM, Chakrader Dewaragatla
>>>> >>> <Chakrader.Dewaragatla@lifelock.com> wrote:
>>>> >>>
>>>> >>> Nifi Team – I would like to understand the advantages of Nifi
>>>> clustering
>>>> >>> setup.
>>>> >>>
>>>> >>> Questions :
>>>> >>>
>>>> >>>  - How does workflow work on multiple nodes ? Does it share
the
>>>> resources
>>>> >>> intra nodes ?
>>>> >>> Lets say I need to pull data 10 1Gig files from S3, how does
work
>>>> load
>>>> >>> distribute  ? Setting concurrent tasks as 5. Does it spew 5
tasks
>>>> per node ?
>>>> >>>
>>>> >>>  - How to “isolate” the processor to the master node (or
one node)?
>>>> >>>
>>>> >>> - Getfile/Putfile processors on cluster setup, does it get/put
on
>>>> primary
>>>> >>> node ? How do I force processor to look in one of the slave
node?
>>>> >>>
>>>> >>> - How can we have a workflow where the input side we want to
receive
>>>> >>> requests (http) and then the rest of the pipeline need to run
in
>>>> parallel on
>>>> >>> all the nodes ?
>>>> >>>
>>>> >>> Thanks,
>>>> >>> -Chakro
>>>> >>>
>>>> >>> ________________________________
>>>> >>> The information contained in this transmission may contain
>>>> privileged and
>>>> >>> confidential information. It is intended only for the use of
the
>>>> person(s)
>>>> >>> named above. If you are not the intended recipient, you are
hereby
>>>> notified
>>>> >>> that any review, dissemination, distribution or duplication
of this
>>>> >>> communication is strictly prohibited. If you are not the intended
>>>> recipient,
>>>> >>> please contact the sender by reply email and destroy all copies
of
>>>> the
>>>> >>> original message.
>>>> >>> ________________________________
>>>> >>>
>>>> >>>
>>>> >>> ________________________________
>>>> >>> The information contained in this transmission may contain
>>>> privileged and
>>>> >>> confidential information. It is intended only for the use of
the
>>>> person(s)
>>>> >>> named above. If you are not the intended recipient, you are
hereby
>>>> notified
>>>> >>> that any review, dissemination, distribution or duplication
of this
>>>> >>> communication is strictly prohibited. If you are not the intended
>>>> recipient,
>>>> >>> please contact the sender by reply email and destroy all copies
of
>>>> the
>>>> >>> original message.
>>>> >>> ________________________________
>>>> >>>
>>>> >>>
>>>> >>> ________________________________
>>>> >>> The information contained in this transmission may contain
>>>> privileged and
>>>> >>> confidential information. It is intended only for the use of
the
>>>> person(s)
>>>> >>> named above. If you are not the intended recipient, you are
hereby
>>>> notified
>>>> >>> that any review, dissemination, distribution or duplication
of this
>>>> >>> communication is strictly prohibited. If you are not the intended
>>>> recipient,
>>>> >>> please contact the sender by reply email and destroy all copies
of
>>>> the
>>>> >>> original message.
>>>> >>> ________________________________
>>>> >>
>>>> >>
>>>> >> ________________________________
>>>> >> The information contained in this transmission may contain
>>>> privileged and
>>>> >> confidential information. It is intended only for the use of the
>>>> person(s)
>>>> >> named above. If you are not the intended recipient, you are hereby
>>>> notified
>>>> >> that any review, dissemination, distribution or duplication of this
>>>> >> communication is strictly prohibited. If you are not the intended
>>>> recipient,
>>>> >> please contact the sender by reply email and destroy all copies
of
>>>> the
>>>> >> original message.
>>>> >> ________________________________
>>>> >
>>>> >
>>>> >
>>>> > --
>>>> > Sent from Gmail Mobile
>>>> > ________________________________
>>>> > The information contained in this transmission may contain privileged
>>>> and
>>>> > confidential information. It is intended only for the use of the
>>>> person(s)
>>>> > named above. If you are not the intended recipient, you are hereby
>>>> notified
>>>> > that any review, dissemination, distribution or duplication of this
>>>> > communication is strictly prohibited. If you are not the intended
>>>> recipient,
>>>> > please contact the sender by reply email and destroy all copies of the
>>>> > original message.
>>>> > ________________________________
>>>>
>>>
>>> ------------------------------
>>> The information contained in this transmission may contain privileged
>>> and confidential information. It is intended only for the use of the
>>> person(s) named above. If you are not the intended recipient, you are
>>> hereby notified that any review, dissemination, distribution or duplication
>>> of this communication is strictly prohibited. If you are not the intended
>>> recipient, please contact the sender by reply email and destroy all copies
>>> of the original message.
>>> ------------------------------
>>>
>>>
>>> ------------------------------
>>> The information contained in this transmission may contain privileged
>>> and confidential information. It is intended only for the use of the
>>> person(s) named above. If you are not the intended recipient, you are
>>> hereby notified that any review, dissemination, distribution or duplication
>>> of this communication is strictly prohibited. If you are not the intended
>>> recipient, please contact the sender by reply email and destroy all copies
>>> of the original message.
>>> ------------------------------
>>>
>>>
>>> ------------------------------
>>> The information contained in this transmission may contain privileged
>>> and confidential information. It is intended only for the use of the
>>> person(s) named above. If you are not the intended recipient, you are
>>> hereby notified that any review, dissemination, distribution or duplication
>>> of this communication is strictly prohibited. If you are not the intended
>>> recipient, please contact the sender by reply email and destroy all copies
>>> of the original message.
>>> ------------------------------
>>>
>>
>>
>

Mime
View raw message