Hi all,
so while I was working on the Kafka Sink I sort of started thinking if such a
thing is a good idea to have at all.
A Kafka system would be able to provide a vast stream of data that would have
to be routed to a PLCs.
On the one side, I haven't seen a single usecase of someone writing to a PLC
with PLC4X and I doubt such a scenario is ideal for the Kafka setup.
A PLC is not able to consume a lot of data until it chokes and stops working.
So in a first round I'll concentrate on implementing a robust Scraper-based
PLC4X Kafka Connect Source, but drop the Sink entirely.
If we find out that a Sink makes sense we'll be able to add it in a later
version.
What do you think?
Chris
Am 02.08.19, 11:24 schrieb "Julian Feinauer" <j.feina...@pragmaticminds.de>:
Hi,
I think that's the best way then.
I agree with your suggestion.
Julian
Von meinem Mobiltelefon gesendet
-------- Ursprüngliche Nachricht --------
Betreff: Re: [KAFKA] Refactoring the Kafka Connect plugin?
Von: Christofer Dutz
An: dev@plc4x.apache.org
Cc:
Hi all,
Ok so it seems that Kafka only supports configuration via simple
unstructured maps.
So if we want some hierarchical configuration like the proposed one, we'll
have to do it log4j.properties-style:
name=plc-source-test
connector.class=org.apache.plc4x.kafka.Plc4xSourceConnector
defaults.default-topic=some/default
sources.machineA.url=s7://1.2.3.4/1/1
sources.machineA.jobs.s7-dashboard.enabled=true
sources.machineA.jobs.s7-heartbeat.enabled=true
sources.machineA.jobs.s7-heartbeat.topic=heartbeat
sources.machineB.url=s7://10.20.30.40/1/1
sources.machineB.topic=heartbeat
sources.machineB.jobs.s7-heartbeat.enabled=true
sources.machineC.url=ads://1.2.3.4.5.6
sources.machineC.topic=heartbeat
sources.machineC.jobs.ads-heartbeat.enabled=true
jobs.s7-dashboard.scrapeRate=500
jobs.s7-dashboard.fields.inputPreasure=%DB.DB1.4:INT
jobs.s7-dashboard.fields.outputPreasure=%Q1:BYT
jobs.s7-dashboard.fields.temperature=%I3:INT
jobs.s7-heartbeat.scrapeRate=1000
jobs.s7-heartbeat.fields.active=%I0.2:BOOL
jobs.ads-heartbeat.scrapeRate=1000
jobs.ads-heartbeat.active=Main.running
Chris
Am 01.08.19, 15:46 schrieb "Christofer Dutz" <christofer.d...@c-ware.de>:
Hi Kai,
that document is exactly the one I'm currently using.
What I'm currently working on is updating the current plugin to not
schedule and handle the connection stuff manually, but use the scraper
component of PLC4X.
Also is the current configuration not production ready and I'll be
working on to make it more easily usable.
But it will definitely not hurt to have some Kafka Pro have a look at
what we did and propose improvements. After all we want the thing to be
rock-solid :-)
Chris
Am 31.07.19, 17:03 schrieb "Kai Wähner" <megachu...@gmail.com>:
Hi Chris,
great that you will work on the connector.
I am not deep technical, but if you need guidance from Kafka Connect
experts, I can connect you to a Confluent colleague to can help
with best
practices for building the connector.
For example, we have implemented a wildcard option into our MQTT
Connector
to map MQTT Topics to Kafka Topics in a more flexible way (e.g.
1000s of
cars from different MQTT Topics can be routed into 1 Kafka Topic).
This
might also be interesting for this connector as you expect to
various PLCs.
This guide might also help:
https://www.confluent.io/wp-content/uploads/Verification-Guide-Confluent-Platform-Connectors-Integrations.pdf
On Wed, Jul 31, 2019 at 4:39 PM Christofer Dutz
<christofer.d...@c-ware.de>
wrote:
> Hi all,
>
> I am currently planning on cleaning up the Kafka Connect adapter
a little
> as this was implemented as part of a proof of concept and is
still I a
> state I wouldn’t use in production ;-)
> But a lot has happened since then and I’m planning on making it a
really
> usable tool in the next few days.
>
> A lot has changed since we created the integration module QT3
2018 and I
> would like to refactor it to use the Scraper for the heavy
lifting.
>
> Currently a user has to provide a parameter “query” which
contains a
> comma-separated list of connection-strings with appended address.
This is
> purely unmanageable.
>
> I would like to make it configurable via JSON or Yaml file.
>
> I think it would make sense to define groups of fields that are
collected
> on one device at an equal rate. So it’s pretty similar to the
scraper
> example, however I would like to not specify the source in the
job, but the
> other way around.
> When specifying the “sources” I would also provide which jobs
should run
> on a given collection.
> As the connector was initially showcased in a scenario where data
had to
> be collected on a big number of PLCs with equal specs,
> I think this is the probably most important use-case and in this
it is
> also probably more common to add new devices to collect standard
data on
> than the other way around.
>
> Also should we provide the means to also set per connection to
which
> kafka-topic the data should be sent to.
> We could provide the means to set a default and make it optional
however.
> When posting to a topic we also need to provide means for
partitioning, so
> I would provide sources with an optional “name”.
> Each message would not only have the data requested, but also the
> source-url, source-name and the job-name with a timestamp.
>
> So I guess it would look something like this:
>
> #
>
----------------------------------------------------------------------------
> # Licensed to the Apache Software Foundation (ASF) under one
> # or more contributor license agreements. See the NOTICE file
> # distributed with this work for additional information
> # regarding copyright ownership. The ASF licenses this file
> # to you under the Apache License, Version 2.0 (the
> # "License"); you may not use this file except in compliance
> # with the License. You may obtain a copy of the License at
> #
> # http://www.apache.org/licenses/LICENSE-2.0
> #
> # Unless required by applicable law or agreed to in writing,
> # software distributed under the License is distributed on an
> # "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
> # KIND, either express or implied. See the License for the
> # specific language governing permissions and limitations
> # under the License.
> #
>
----------------------------------------------------------------------------
> ---
> # Defaults used throughout all collections
> defaults:
> # If not specified, all data goes to this topic (optional)
> default-topic: some/default
>
> # Defines connections to PLCs
> sources:
> # Connection to a S7 device
> - name: machineA
> # PLC4X connection URL
> url: s7://1.2.3.4/1/1
> jobs:
> # Just references the job "s7-dashboard". All data will be
published
> to the default topic
> - name: s7-dashboard
> # References the job "s7-heartbeat", however is configures
the
> output to go to the topic "heartbeat"
> - name: s7-heartbeat
> topic: heartbeat
>
> # Connection to a second S7 device
> - name: machineB
> url: s7://10.20.30.40/1/1
> # Sets the default topic for this connection. All jobs data
will go to
> "heartbeat"
> topic: heartbeat
> jobs:
> - s7-heartbeat
>
> # Connection to a Beckhoff device
> - name: machineC
> url: ads://1.2.3.4.5.6
> topic: heartbeat
> jobs:
> - ads-heartbeat
>
> # Defines what should be collected how often
> jobs:
> # Defines a job to collect a set of fields on s7 devices every
500ms
> - name: s7-dashboard
> scrapeRate: 500
> fields:
> # The key will be used in the Kafka message to identify
this field,
> the value here contains the PLC4X address
> inputPreasure: %DB.DB1.4:INT
> outputPreasure: %Q1:BYTE
> temperature: %I3:INT
>
> # Defines a second job to collect a set of fields on s7 devices
every
> 1000ms
> - name: s7-heartbeat
> scrapeRate: 1000
> fields:
> active: %I0.2:BOOL
>
> # Defines a third job that collects data on Beckhoff devices
> - name: ads-heartbeat
> scrapeRate: 1000
> fields:
> active: Main.running
>
> I think it should be self-explanatory with my comments inline.
>
> What do you think?
>
> Chris
>
>
