OpenSCADA

[gtludwig]

is there any detailed how to for the installation of OpenSCADA latest
version?

[ctron]

Hi,

well we have start at http://download.openscada.org/documentation/I/0.15.0/doc/
but it still needs a lot of work. But you are welcome to help ;-)

[Gustavo Tomazi Ludwig]

Hi,

I have those docs. =)

I am involved in a opensource SCADA project as a developer. We (my team
and I) have tried to use your software last December and were frustrated
by its not-so-easy installation. Thought your software looked promising
(as to stability and reliability and more), we had a project milestone
ahead and we started to develop on Serotonin's Mango M2M.

Our main contributions for this projects so far were the implementation
of an external API so data can be consumed as Web Services using SOAP,
adding a DNP3 protocol, Brazilian Portuguese internationalization and a
few other tweaks here and there.

So, from an "advanced user" point of view, I'd like to run OpenSCADA to
evaluate it since we're considering changing our DA server to a more
stable, fast, secure and redundant one. From what I could fathom,
OpenSCADA seems to do this (and more!).

I understand OpenSCADA is divided into several modules and as far as I
could tell, it has some dependencies and a correct install order, right?

From what I understood, I need to install in this order:

1) External;
2) Aurora;
3) Atlantis;
4) Utgard, ...

In exchange for your help, I'll gladly format my steps to get OpenSCADA
running in a way I hope will be of use for your project, my team and
(hopefully!) developer community.

Thanks in advance,

Gustavo

[Jens Reimann]

Well that sounds good ;-) We are a little bit busy at the moment. So I don't have too much time left.

Installing depends on what you like to do. Developing for OpenSCADA or using OpenSCADA. If you want to use it it depends on what you would like to use. The client can be easily downloaded from  http://download.openscada.org/orilla/I/0.15.0/ostc/ simply extract and run the test client and the test server. The two addresses that can be used when the test server is started from the "DA" perspective "testing" menu is "da:net://localhost:1202" and "da:net://localhost:1203" The port numbers can be changed in the preferences dialog.

If you want to install some simple "DA servers" you would need to set up a directory containing all required (or all) JARs and launch die "Exporter". This all is quite a bit complicated but if you like and if I find the time I would document that for you.

On the other hand if you want to use the "full feature super server" including AE and HD you need to set up an OSGi container (most easily Equinox from inside Eclipse) and start all the bundles you like. If you have experience with Eclipse and Equinox this shouldn't be a big problem. Just add the repositories to your target platform (should be described in the documentation) and set up a a Run Configuration. You should start the "org.openscada.ca.file" and "org.openscada.ca.servlet" bundle and the jetty webserver bundles so you can use the web configuration at "http://localhost:8080/ca" (or whatever jetty port you are using.

After you got that running you got a server doing nothing ;) But you can start configuring stuff in the "master server". The problem is that you don't know that to configure. But then again I could also document a quick example for you. In the documentation we started to document all the OSGi bundles and configurable services we have. But this will take some time until it is complete. But a quick and simple sample should not be a problem.

Now I hope I haven't confused you too much ;) The easiest start would be to run the OSTC and try out the embedded test servers.

greetings

jens

--
To unsubscribe, reply using "remove me" as the subject.

[Gustavo T. Ludwig]

How can I checkout OpenSCADA 0.15.0 from svn?

[Andrey Kogun]

http://pubsvn.inavare.net/openscada/modules/

--
Андрей

[Jens Reimann]

I would suggest you check out everything directly in eclipse 3.5:

http://pubsvn.inavare.net/openscada/modules/{atlantis,aurora,external,orilla,oxygen,utgard,deploy}/trunk/* ... you get the pattern ;)

next you create a new target platform and add a plain "eclipse_home" as installation and append a "directory" to "${workspace_loc:/builder_external/builder/lib}" (guess the syntax should be right.

activate the target platform and let it refresh.

this will give you all there is for your purpose.

[ctron]

Hi,

I used some pauses today and tried to add a "quick start" section in
the documentation. See http://download.openscada.org/documentation/I/0.15.0/doc/html/userguide/index.html#quickStart

Although I have to admit that I did not test it at all. I gathered
everything from my local test projects ... so hopefully it helps you a
little bit.

Btw ... one common thing that is forgotten when checking out openscada
is that one has to run some ant scripts in some projects. some
projects have "custom-build.xml" files which are run by the PDE build.
But they only trigger some other ant script in the same project. You
have to start "the other" file. They should be named "support-
build.xml" in most cases. This is a bit ugly but works quite well for
PDE and IDE builds. The one other thing you need is "slice2java" from
the "ICE" project of ZeroC. The need some tool to convert their IDL
files into sourcecode. I am not sure if you can simply close the
"*ice*" projects (you can try and won't miss a thing) but there might
be dependencies you need. This is always a bummer and already
considered dropping support for ICE since it makes more pain than it
heals.

Anyway you won't need ICE and can simply kick it out.

I am off for the Weekend now. If you got any questions just mail them
and maybe if have a look. But I won't promise it ;-)

On 16 Apr., 17:13, "Gustavo T. Ludwig" <gtl.li...@gmail.com> wrote:
> How can I checkout OpenSCADA 0.15.0 from svn?
>
> On 04/16/2010 04:56 AM, Jens Reimann wrote:
>
> > Well that sounds good ;-) We are a little bit busy at the moment. So I
> > don't have too much time left.
>
> > Installing depends on what you like to do. Developing for OpenSCADA or
> > using OpenSCADA. If you want to use it it depends on what you would
> > like to use. The client can be easily downloaded from

> >http://download.openscada.org/orilla/I/0.15.0/ostc/simply extract and

[Gustavo T. Ludwig]

Hi,

I managed to svn checkout trunk directories for Atlantis, Aurora, Deploy, External, Orilla, Oxigen and Utgard modules into a fresh Eclipse 3.5 install.

But when I try to run External/builder_external/build.xml, Ant returns this message:

I have OpenSCADA in /home/<user>/java/OpenSCADA/, which I'll call OS_HOME now on.

in ON_HOME I have:
eclipse/
workspace/ (all modules are downloaded here)
ostc/

when I cd workspace/External/builder_external and try

ant build.xml

I get this:

BUILD FAILED
/home/gtl/java/OpenSCADA/workspace/Atlantis/builder_atlantis/build.xml:38: The following error occurred while executing this line:
/home/gtl/java/OpenSCADA/workspace/Atlantis/builder_atlantis/common/lib/lib.xml:83: /home/user/eclipse/plugins/org.eclipse.pde.build_3.5.0.v20090527-1800/scripts is not a valid directory

seems a pde related issue, even though on OS_HOME/eclipse/plugins I have these:
org.eclipse.pde.build_3.5.1.R35x_20090820/
org.eclipse.pde.build_3.5.2.R35x_20100114/

Any ideas?

Thanks in advance,
gtludwig

[ctron]

Hi,

yes ... you need to adapt the "build.properties" file or, even better,
create a file named "local.properties" beside it. It is located in the
"builder_*" directories at "common/config". all properties in this
file override the build.properties. the following should be sufficient
for Eclipse 3.5.2

eclipse.home.dir=/home/user/some/path/to/eclipse
eclipse.launcher.version=_1.0.201.R35x_v20090715
eclipse.builder.version=_3.5.2.R35x_20100114

the following line reduces the build to only produce linux/gtk/64 bit:

configs=linux, gtk, x86_64

But you should need all that only if you compile OpenSCADA using the
headless build system. If you check out into eclipse you won't need to
build from the command line.

I also added some more documentation and a sample project for one of
the the "non OSGi" driver application (the "shell script server",
quite useful). It is described at
http://download.openscada.org/documentation/I/0.15.0/doc/html/userguide/ch01s02.html#quickStart_exec

I hope this helped you a little bit.

greetings

jens

> > the documentation. Seehttp://download.openscada.org/documentation/I/0.15.0/doc/html/usergui...

> >>>http://download.openscada.org/orilla/I/0.15.0/ostc/simplyextract and

[Gustavo T. Ludwig]

Hi (again!),

Which Eclipse distribution is most recommended? I just got this one:

eclipse-jee-galileo-SR2-linux-gtk-x86_64.tar.gz

Both ANT_HOME and M2_HOME (Maven2) are exported as system variables.

I did checked out into Eclipse.

I still don't get what should I run and on what order...

BTW, sorry such noobish questions =)

Thanks,
gtludwig

[Gustavo T. Ludwig]

actually, would anyone be available for a quick help-desk session over at Skype?

On 04/19/2010 02:18 PM, ctron wrote:

[Jens Reimann]

Hi, eclipse 3.5.2 is what we currently use. But any 3.5 should be fine if you stick to the IDE.

The quick start tutorials in the new documentation I posted the links to should give you some help. The osgi based example also list the bundles you need for a really simple sever.

The quick start sample for the exec server should give you something that can be start from the command line really quick.

  

On Apr 19, 2010 7:44 PM, "Gustavo T. Ludwig" <gtl....@gmail.com> wrote:

Hi (again!),

Which Eclipse distribution is most recommended? I just got this one:

eclipse-jee-galileo-SR2-linux-gtk-x86_64.tar.gz

Both ANT_HOME and M2_HOME (Maven2) are exported as system variables.

I did checked out into Eclipse.

I still don't get what should I run and on what order...

BTW, sorry such noobish questions =)

Thanks,
gtludwig

On 04/19/2010 02:18 PM, ctron wrote: > > Hi, > > yes ... you need to adapt the "build.properties" f...

[Andrey Kogun]

Hi,

Thank you for your help. I was able to run the OSGi based master server and DA client from the example.

Now the challenge to make a few servers for data collection from various sources and recording the data in the database (most likely Postgree). Please advise what can be taken as the basis for these tasks? 

Another question, where you can see a list of urgent tasks for OpenSCADA project. 
Is it possible to take part in development process?

--
Андрей

[ctron]

Congratulations! :)

So you still might check out the other "exec server" tutorial before
you decide on how and what to write for DA.

At the moment we have two types of DA server. Those based on OSGi and
those before OSGi. While they have a lot in common (from the source
code side of view) they are pretty different in matters of
configuration.

The OSGi stuff is designed to handle configuration updates dynamically
during runtime and is based on a key/value based storage.

The non-OSGi stuff takes its configuration from "somewhere" (mostly
XML files, based on Apache XML Beans) and only loads its configuration
on startup.

Also the non-OSGi stuff is a little bit older but more mature while
the OSGi stuff has provides some features the non-OSGi has not.

Maybe you should have a look at some other drivers that already exist
first. At the moment we have the following "out of the box"
interfaces:

- Exec: Shell script / command line stuff
- JDBC: JDBC Queries
- Proxy: Aggregate one ore more sources to one. for proxying or
redundancy switch over
- OPC: OPC DA 2.0
- SNMP
- S7 PLC
- some testing servers like (Spring Framework, Testing, Sysinfo and
Stock)

The S7 PLC is based on OSGi. All others are "plain Java" servers which
can also be started using the Apache Daemon framework.

We got a couple more servers, but I cannot provide you the source code
for these since they are specifically developed for our customers.
Only the Modbus (TCP and RTU) driver is currently "undecided" and I
hope I will get it into OpenSCADA sooner or later.

So I would suggest you check out the Exec Server at least to gain a
feeling for "non-OSGi" stuff before you decide which way to go.

If you are curious you can try and check any of the "datasource"
bundles. Make a copy and implement a data source based on
"org.openscada.da.datasource.base.AbstractDataSourceHandler". The
difference between datasource and data item is that a data item is a
data source in a DA Hive that provides one value and allows
differntial transmission of data. While the datasource is a based on
the idea of a data item but reduces the complexity by always
transmitting the full status. That is why datasources are using
internally inside of the OSGi container and the data items are used on
the external interfaces. The non-OSGi stuff uses data items internally
and externally where the OSGi stuff uses externall data items (due to
the interface) and internally data sources.

So creating your own bundle with your own data source implementation
provides you with a datasource that can be exported as data item. On
the other side you can also implement some class that simply attaches
to a datasource (or data item) and simply writes the data to a
database.

then again we have something called "hd server" which are some
components that are available for the OSGi server (org.openscada.hd.*)
which provide you with a historical database. Sadly this piece of
software is not available in the quality it should be so it has to be
handled with extreme care and I don't want to throw you into water
that is too deep ;-)

On Apr 19, 9:12 pm, Andrey Kogun <a.ko...@gmail.com> wrote:
> Hi,
>
> Thank you for your help. I was able to run the OSGi based master server and
> DA client from the example.
>
> Now the challenge to make a few servers for data collection from various
> sources and recording the data in the database (most likely Postgree). Please
> advise what can be taken as the basis for these tasks?
>
> Another question, where you can see a list of urgent tasks for OpenSCADA
> project.
> Is it possible to take part in development process?
>
>
>
> On Mon, Apr 19, 2010 at 10:05 PM, Jens Reimann <ctron1...@gmail.com> wrote:
> > Hi, eclipse 3.5.2 is what we currently use. But any 3.5 should be fine if
> > you stick to the IDE.
>
> > The quick start tutorials in the new documentation I posted the links to
> > should give you some help. The osgi based example also list the bundles you
> > need for a really simple sever.
>
> > The quick start sample for the exec server should give you something that
> > can be start from the command line really quick.
>

> > On Apr 19, 2010 7:44 PM, "Gustavo T. Ludwig" <gtl.li...@gmail.com> wrote:
>
> >  Hi (again!),
>
> > Which Eclipse distribution is most recommended? I just got this one:
>
> > eclipse-jee-galileo-SR2-linux-gtk-x86_64.tar.gz
>
> > Both ANT_HOME and M2_HOME (Maven2) are exported as system variables.
>
> > I did checked out into Eclipse.
>
> > I still don't get what should I run and on what order...
>
> > BTW, sorry such noobish questions =)
>
> > Thanks,
> > gtludwig
>
> >  On 04/19/2010 02:18 PM, ctron wrote: > > Hi, > > yes ... you need to adapt
> > the "build.properties" f...
>
> --
> Андрей
>

> --
> Subscription settings:http://groups.google.com/group/openscada/subscribe?hl=en

[Gustavo T. Ludwig]

Hi all,

I managed to get the OSGi server running!

I do:

(...)

Add the following command line arguments for the Java™ virtual machine: -Dorg.eclipse.equinox.http.jetty.http.port=8082 -Dopenscada.da.net.exportUri=da:net://0.0.0.0:1204 -Dorg.openscada.ca.file.root=/home/user/cas/test

Now start the configuration and wait for the OSGi™ prompt.

1.2.2.3. Configure the master server

Now head your web server to http://localhost:8080/ca and create the following configuration objects:

• da.connection with id connection.1

  • connection.uri=1203

• da.datasource.dataitem with id datasource.1

  • connection.id=connection.1

  • item.id=memory-cell-0

• da.dataitem.datasource with any id

  • datasource.id=datasource.1

  • item.id=alias.item.1

1.2.2.4. Connecting

Start the OSTC and connect to da:net://localhost:1204. You should find an item named alias.item.1 which (when draged to the real time list) is “CONNECTED” but “NULL”.

Start the local test servers from inside OSTC and write (e.g.) 5 to the memory cell control item as described in Section 1.2.1, “Test Client”. Now write some value on the item memory-cell-0.

After the write operation is complete you should see that also the item alias.item.1 has changed to the value you just wrote.

But when I connect on OSTC, there are no item named alias.item.1.

Questions:
1) shouldn't ports be the same on Java VM args on item 1.2.2.2 and URL on item 1.2.2.3?
2) shouldn't 'connection.uri=1203' be 'connection.uri=da:net://0.0.0.0:1204' according to Java VM args?
3) there should be a 'connection.id= connection.1' at da.datasource.item too, right?

I've attached a screenshot of http://localhost:8082/ca (48kb) and of the connected datasource on OSTC (43kb).

What am I missing now?

Thanks in advance!

Gustavo

[Jens Reimann]

Hi,

well I am not sure are the moment and a bit busy to try it out. But I guess that the configuration parameters for "da.dataitem.datasource" and "da.datasource.dataitem" have to be swapped. One requirement at the moment is that the exported data item has at least one "." (dot) in the name otherwise it is not shown in the browser.

1) shouldn't ports be the same on Java VM args on item 1.2.2.2 and URL on item 1.2.2.3?
2) shouldn't 'connection.uri=1203' be 'connection.uri=da:net://0.0.0.0:1204' according to Java VM args?
3) there should be a 'connection.id= connection.1' at da.datasource.item too, right?

for these three items: there are two servers running in this example. The test server from OSTC on port 1203 and the master server on port 1204. The datasource that is exported by the master server does not need a connection uri/id since it is exported the the application default exporter.

[Gustavo T. Ludwig]

ok, I got it working properly... I guess! :)

I'm interested now in protocols. Which are currently implemented?

Thanks in advance,
Gustavo

[Gustavo T. Ludwig]

Hi all,

How can I simulate redundancy with OpenSCADA?

As far as I understand, redundancy means a "backup" datasource. In
OpenSCADA terms, only one dataitem attached to more than one
datasources, so if one datasource fails, the next one automatically
kicks in.

Is this possible? How so?

Thanks in advance,
Gustavo

[Jens Reimann]

Hi, today is not a good day for answers. I guess I will be back on the office tomorrow. But I asked my collegue to answer your question. So if he has the time he will explain.

[Jürgen Rose]

Hi,

the class org.openscada.da.datasource.proxy.internal.ProxyDataSource
implements the redundancy on datasource level. It essentially receives
updates from all sources and notifies its listeners if the value
actually has changed. So if it receives the same value from all of its
sources it is propagated only once. Actually the logic is more allong
the following: 1.) a handler receives a update 2.) the update() method
of the ProxyDataSource is called. 3.) in this method all handlers get
sorted by priority and the "best" one is selected (previously all
disconnected handlers are discarded) 4.) updateData() of the base
class will be called if value has actually changed, which propagates
it to all subscribers

I attached a picture how the configuration would look like, although
realistically there would be different connections for each source
item (memory-cell-0 and memory-cell-1).

The current implementation doesn't allow to write values back (it will
be some time in the future).

There is also a second (older) possibility to achieve redundancy:
using the proxy hive (org.openscada.da.server.proxy.Hive). This is a
give which serves as a gateway to potentially multiple other hives. It
essentially connects as a client to the other OpenSCADA servers and
proxies their items (here values may also be written!). But here the
fallback connection has to be explicitely set (this may also have
advantages in certain situations). We used this in a customer project,
together with a second component which watched the master flag of the
two PLC servers, and if the master flag changed (or connection was
lost), the underlying connection of the proxy hive was switched as
well.

Jürgen

2010/4/26 Gustavo T. Ludwig <gtl....@gmail.com>:

[Gustavo T. Ludwig]

Thanks! =)
Trying this now!

[Gustavo T. Ludwig]

Hi,

I'm running OSTC in a computer at 150.162.165.148:1203 where I started a Local Test Server. On this server I have 2 memory-cells, onto which I started a signal generator. Configuration for both is:
Minimun: -1;
Maximun: 1;
Period: 2000;
Function: Sine.

On another computer, I have OSTC connected to the Server and values for memory-cell-0 and memory-cell-1 are generated on the server and displayed on the client.

Now on my client computer, I start a server that connects to 150.162.165.148:1203 and tries to implement redundancy, but fails...

Here's what I have done:

da.connection:
connection.uri=da\:net\://127.0.0.1\:1204
id=connection.1

connection.uri=da\:net\://150.162.165.148\:1203
id=connection.2

connection.uri=da\:net\://150.162.165.148\:1203
id=connection.3

da.dataitem.datasource:
item.id=alias.redundancy.1
datasource.id=proxy.1
id=dataitem.0

da.datasource.dataitem:
connection.id=connection.2
item.id=memory-cell-0
id=datasource.2

connection.id=connection.3
item.id=memory-cell-1
id=datasource.3

da.datasource.proxy:
sources=datasource.2,datasource.3
id=proxy.1

If I stop signal generator for memory-cell-1, alias.redundancy.1 gets its value from memory-cell-0. But if I stop signal generator for memory-cell-0, alias.redundancy.1 also stops.

I intend alias.redundancy.1 gets it's value from whatever valid value.

I did not understood your second way.

What am I doing wrong here?

Thanks in advance!
gtludwig

On 04/26/2010 01:55 PM, Jürgen Rose wrote:

[Gustavo T. Ludwig]

Hi,

Got it working! I was looking at the wrong class. =)

Thanks!
Gustavo

On 04/26/2010 01:55 PM, Jürgen Rose wrote:

[CptMauli]

Great, I just was analyzing what could have gone wrong :-)

Jürgen

On 30 Apr., 14:42, "Gustavo T. Ludwig" <gtl.li...@gmail.com> wrote:
> Hi,
>
> Got it working! I was looking at the wrong class. =)
>
> Thanks!
> Gustavo
>

> > J�rgen
>
> > 2010/4/26 Gustavo T. Ludwig<gtl.li...@gmail.com>:

[Gustavo T. Ludwig]

Hello all,

I'd like to build a dist package of Utgard complete with all sources and
Javadocs to use it as a lib in our application.

I know that I can find the latest Utgard here:
http://openscada.org/download/maven-repo/org/openscada/utgard/org.openscada.utgard/0.5.0-SNAPSHOT/

And its dependencies here:
http://openscada.org/download/maven-repo/org/openscada/utgard/org.openscada.utgard.opc.dcom/0.5.0-SNAPSHOT/
and here:
http://openscada.org/download/maven-repo/org/openscada/utgard/org.openscada.utgard.opc.lib/0.5.0-SNAPSHOT/

All directories have the package, javadoc and sources jars and a file
that binds them all called <package-name>.pom

OK, I understand you don't use maven any longer and it may be that you
can't offer me support on this issue, but if you can throw me some
pointers it would be most appreciatted!

Thanks in advance,
Gustavo

[ctron]

Hello Gustavo,

well ... all the Maven stuff is deprecated and no new modules are
provided. We don't use maven anymore since we got too much headaches
from it ;-)

There is a P2 repository at:
http://download.openscada.org/utgard/I/0.7.0/org.openscada.utgard.sdk-0.7.0-p2repo.zip
And a plain ZIP archive at: http://download.openscada.org/utgard/I/0.7.0/openscada-utgard-I.0.7.0.zip
contain the binaries

I guess it would be possible to add the sources too.

If you really want to use maven, you could use our P2 to maven
converter. Which converts you some P2 repositories to a maven
repository.

best regards

jens

On 27 Mai, 15:35, "Gustavo T. Ludwig" <gtl.li...@gmail.com> wrote:
> Hello all,
>
> I'd like to build a dist package of Utgard complete with all sources and
> Javadocs to use it as a lib in our application.
>

> I know that I can find the latest Utgard here:http://openscada.org/download/maven-repo/org/openscada/utgard/org.ope...
>
> And its dependencies here:http://openscada.org/download/maven-repo/org/openscada/utgard/org.ope...
> and here:http://openscada.org/download/maven-repo/org/openscada/utgard/org.ope...

[Parandhaman GR]

Hi, I just tested the OSTC works good with test servers.

Can I use this test server to connect to OPC DA server example (Matrikon or KEPServer)?

If yes, what will be the connection url say for eample the KEPServer is running in the same PC with few channels and devices.

Kindly help me out I need an open source OPC Client.

[CptMauli]

Hi,

unfortunately not, the OSTC only supports the Eclipse SCADA internal protocol. This actually might be an interesting feature for a future version to integrate the OPC adapter with the OSTC.

What you can do is to run the OPC driver separately which connects to your OPC servers and then access that one with the OSTC.

best regards
Jürgen