Headstart and Client/Server (C/S)

Hi 'Generators',
I have a question related to running Headstart generated Forms within a C/S environment. Is it possible to get a stable C/S environment?
I have checked the Headstart installation manual and there is a chapter "Using Headstart 6.5 in a Client/Server Environment". I have implemented the changes to the object library (where can i find "progress_indicator.amount" by the way???) but the forms are still not very stable in the client server mode. When i run the form in Oracle forms developer standalone there are no problems. When i run the form from the start form it crashes almost immediately. HEEEEEEEEEEEEEEEEEELLLLLUUUPPPPPP!!!!! Have i missed an undocumented change to the object library????
If somebody was able to create a stable C/S environment!!! Please let me know!!!!
Gr.
Louk

Hi Lauri,
First of all, thanks for your information and your e-mail.
This is what i have tried at the beginning of this week. To see if text "automatic" was still in my forms i converted my forms from FMB to FMT format and did a search and replace on "automatic". For every text "automatic" found i have changed it to some kind of default color and then converted the forms back to FMB and then i have tried them to run again. The forms still crash. The strange thing is that they run correct within the FORM builder as stand-alone forms (open form in forms builder and then hit the traffic light to run the form). When i run them within the main form (start-form with menu's to call my forms) things start to go wrong. What is the difference? Is there something i forget? Can you tell me if there is someone out there who was able to get a stable environment at all?
Our plan to run the forms in a C/S environment using Citrix seems to be over when it is not possible to get it stable. We have to use the WEBFORMS environment instead and this means we have to invest in the Forms Server. Or are there other alternatives?
Gr.
Louk

Similar Messages

Timezones and client/server issue

We have a client process which runs on Windows PCs which are scattered among different timezones. These clients send data periodically to a server.
The data includes timestamps indicating when certain events happened on each client PC. We need to know the timezone for each PC so that we can display those timestamps (stored on the server in GMT) in the timezone of each PC.
The tricky bit: the clients are in C++ and the server is in Java. We cannot assume a JRE is installed on the clients.
Our C++ client code can get the Windows timezone name as a string and pass that to our Java server code. What isn't obvious is how to translate that string to a Java TimeZone object. For example, the Windows string "Central Standard Time" maps to the Java TimeZone ID String "America/Chicago".
I've done enough research to discover that the (relatively new) jre/lib/tzmappings file contains exactly these mappings. But it appears that the code which relies on that file is quite "under the hood" in java -- most likely in the unpublished sun.util.calendar package.
Does anyone know of a way to map an O/S-specific timezone string to a Java TimeZone object? I know I can write my own code to read the tzmappings file myself, but that file explicitly states that "No applications should depend on this file in any form." I'd prefer a kosher solution.
All ideas, suggestions, thoughts, links, etc. appreciated. Thank you!

Thanks for the comments, everyone. Some responses:
It seems the exact opposite to me. Relying on yourown file rather
than using a file that explictly says "Don't useme!" seems less brittle.
Both are brittle. By relying on the "one built into
Java" I can rely on timezone additions and changes
made by Sun without doing anything. Otherwise I have
to be responsible for maintaining mine over time,
noticing (for example) whenever Sun updates theirs
and then updating and re-releasing mine.
You either are going to be mixing data or not.Can you clarify for me what you mean by "mixing
data"?
We were planning to store all timestamps in the
database in GMT with a timezone code. This allows
queries to be based on normalized time (all use GMT),
yet also allows us to know when an event happened in
"local time" (on the PC where the event occurred) by
converting the GMT value using the timezone code.There is a difference between doing a query and displaying the data.
In terms of the query you don't need to access anything (special files.) You do the query in SQL, nothing more.
In terms of displaying the data you have a number of choices.
1. Don't display the time at all.
2. Display it in reference to a specific timezone.
3. Display each time in reference to the timezone that originated it.
The last choice is not usable for users when the data is mixed in one display.
>
it is going to be a very large mistake to attemptto display each time in
a different timezone (without categorizing it someother way.) It will be
non-sensical to the users.What about when the clients ARE in various timezones?
We have GUI screens and reports which show events on
PCs in various continents. For example, showing:
Again you are going to be doing one of the following.
1. Data that is relative to them only.
2. Data from all zones.
If the first then just display the time relative to them in the first place (use the default for the VM.)
If the second then you are back to the problem that I discussed above.
MachineA 09:01 Service Started
MachineA 10:32 Service Stopped
MachineB 10:33 Service Stopped
makes it look like the events (Service Stopped)
happened about a minute apart on MachineA and
MachineB. But they didn't, in this example, because
the machines are in two different timezones. A user
can't tell that above. But this clarifies it:
MachineA 09:01 PST Service Started
MachineA 10:32 PST Service Stopped
MachineB 10:33 EST Service StoppedNote that I am not suggesting that you do not store your data in an absolute sense.
But storage and display are two different things. And your example is flawed unless your database data is flawed.
For example if you store the time in the database as GMT then the only way you can get case 1 is if you use the local timezone for each and then don't display the timezone.
>
In fact, the service stopped on MachineB before
MachineA.
The question begging to be asked is: what does the
user expect to see? A list of events sorted relative
to "wall time" (shown above), or a list sorted by
absolute time? The latter looks like this:I think you will find that most users expect time to be relative to a certain timezone when times from different timezones are mixed in the same display.
If grouped by timezone then they will accept timezone relative times even when there are different view with different timezones.

JSP and Client/Server Compatibility Issues

For the new application that we have to design we have two groups of users
- Group A requires admin access, high response time, better user feel and special previledges and are a few in number at limited locations.
- Group B requires limited functionality, but are spread all over the world and they are huge in number.
We are planning to provide Applet/Swing based client server architecture for Group A and JSP/EJB based J2EE architecture for group B users. common webserver would be used.
Please advise what are are potential issues in this approach or do we have a better approach to achieve the same.

GUI based applications are good but do you think that will it be maintainable and portable?.. Some of the GUI based applications are build on higher Java like 1.5.. i Presume you have the latest and other workstation have only 1.4 or 1.3, this can be n issue on GUI based applications especially on installation on each workstation.

Uml and client server programming

I m developing chat messenger. i am finding difficult to apply uml on this type of software (client server).can any one give me sample documentation or tutorials

UML can express and communicate your
ideas for sure, but it's not where to start with the
high level architecture unless you have an idea about
how to achieve it.
Once you have an idea of how it should work, then the
activity diagrams/sequence diagrams/collaboration
diagrams will just drop out of you and into the model.Actually, UML usually is the first place to start. Specifically, the Use Case diagram is often the starting point in designing a system. The first iteration of these diagrams will be very rough for sure, but what better way to begin discovering the functionality of a system than to (at a very high level) determine who's gonna use the system and how they are going to use the system.
A common practice is to get a sheet of paper and draw a line down the middle (making two column). Label them actors (left colum) and processes (right column). Starting with the right column, list (again, at a very hight level) the thins that your system should do. Repeat this for the left column, this time thinking of the "types" of users that will use the processes you've just listed. Then you draw lines from the actors to the processes and this is the begining of your use cases. The next steps is to start storyboarding the use cases. This forces you to think of what each process does in in a little more detail. From there, you will begin to see common behavior and even new use cases. It takes a few iterations to fully discover and realize use cases, but you start with the 10,000' view of it this way.
For example (and this is just an example):
actors               processes
System               Post a message
Authenticated User   Read a message
Anonymous User       Delete a message thread
Administrator        Login
                     Logout
                     Register
                     Update Profile
                     Generate Audit EntryWhen you draw the lines from actors to processes in the above example, you'll note some the following:
1. System can Generate Audit Entry
2. Authenticated User can Post a message, Read a message, Logout and Update Profile
3. Anonymous User can Reade a message, Login and Register
4. Administrator can Post a message, Read a message, Delete a message thread, Logout and Update Profile
Again, this is just an example. You should see that it give you a bird's eye view of the services to be provided by your system and who can use those services. When you begin realizing your use cases (the processes in the above example), you will begin to see behavior that is common to many services. These behaviors are what will begin helping you to eventually make architecture level decisions.

PDK and Client/Server

Hello,
is it possible to create a portlet for a normal client/server application like MS Word oder PowerPoint to be displayed/used in Oracle Portal? Does the PDK solve the problem or do i need web-enabled versions of these programms?
If it is possible do so, could you please give me a shot discription of how to do it.
greeting
Thomas

GUI based applications are good but do you think that will it be maintainable and portable?.. Some of the GUI based applications are build on higher Java like 1.5.. i Presume you have the latest and other workstation have only 1.4 or 1.3, this can be n issue on GUI based applications especially on installation on each workstation.

Socket Communication between java client and c++ server

HI,
In my project,I want to do the following:
1.Sending datas from client to server.
2.Getting the response from server to client.
I written the client in java.but the server is in c++.
Is it possible to communicate with the server using java codings itself?
Im able to send the data from my java client to the server.
but unable to get back the datas from server to client.
Can anyone tell me how to do this?
thanks a lot

hi
thanks for ur reply.
I didnt get any error msg while getting the back the datas.
Actually i divided my application into two parts.
My application will act as both server and client.
server ll get the browser request and send to the client and the client will send that data to the c++ server.
Im able to do that.and unable to get the data from server.
Didnt get any error.
can u tell me how to make an application to act as both client and server.
I think im wrong in that part.
thanks a lot

Client Server mode for time ten connection

Hi, I am new to Times Ten and got a question for running TimesTen in Client Server Mode. When it is running in Client/Server mode:
Does the database just run like a regular remote database ?
Is there security for the remote application, such as userid/password ?
Is there any limitation on number of remote connection can concurrently connect to the DB ?
Does it use the JDBC or it has to be the native library ?
Thank you for your help
Henry

Hi,
With regard to your questions regarding TimesTen client/server mode:
1. Does the database just run like a regular remote database ?
I'm not sure exactly what you are getting at here but yes, I would say that it runs like any other c/s database. If you can be more specific in your question then I can give a more precise answer.
2. Is there security for the remote application, such as userid/password ?
If the TimesTen instance to which you are connecting was installed with Access Control enabled then yes, normal userid/password security applies. The client application must provide a valid TimesTen userid and password in order to establish a connection.
3. Is there any limitation on number of remote connection can concurrently connect to the DB ?
Subject to O/S resource limitations, a single instance of TimesTen (i.e. the TT server) supports a maximum of 9,999 concurrent client/server connections. Each individual datastore supports a maximum of 2047 concurrent connections (direct and client/server).
4. Does it use the JDBC or it has to be the native library ?
The TimesTen JDBC driver supports both direct modde and client/server. The TimesTen ODBC driver is provided in both direct mode and client/server versions. The JDBC driver uses the ODBC driver (it is a type 2 JDBC driver) and so the client machine requires a TimesTen client installation to use either ODBC or JDBC to access a remote datastore.
Chris

Why timesten is slow in client/server mode

i am testing Timesten client/server mode and find that
it is to slow.
when using dircet mode, timesten can precoss about 40000 query pre second
but when change to client/ server mode
it can only process 1500 query pre second.
here is my config and test sql.
[Test_tt702]
Driver=//u01/oracle/TimesTen/tt70/lib/libtten.a
DataStore=//u01/oracle/TimesTen/tt70/info/DemoDataStore/Test_tt702
DatabaseCharacterSet=US7ASCII
TempSize=64
PermSize=250
Temporary=1
TypeMode=1
CREATE TABLE TEST999
A1 number NOT NULL PRIMARY KEY ,
A2 number,
A3 number
network card is 1000m,
query code:
PreparedStatement pSel = dbConn.prepareStatement("select a1,a2 from TEST999 where a1=?");
rs = null;
int a1=0;
int a2=0;
for(int i=0;i<80000 ;i++)
     pSel.setInt(1,i );
rs=pSel.executeQuery( );
if (rs.next()) {
     a1 = rs.getInt(1);
     a2 = rs.getInt(2);
     if( i%8000==0) {
     System.out.println("select "+ i +" res="+a1 +" time " + getTime_v2());
     rs.close();
direct mode time : 80000 query is 2 second
client/server mode in the same machine : 80000 query is 12 second
client/server mode in the different machine : 80000 query is 54 second
any one have idea about this?
is it the jdbc driver 's promble?

it is to be expected that there will be a big difference in performance between direct mode and client/server due to much greater overhead in client/server, especially if the client is on a different machine. However, the differences you are seeing here are larger than I would normally expect.
You say (I think) that the network is 1 GB, correct? What is the hardware spec of the test machine(s)? Have you tuned the O/S network stack for optimal performance?
Typically, for local client/server using the fastest IPC mode (shmipc) I expect performance of around 20-30% of direct mode and for remote client/server with a GB LAN I would expect performance of around 10-20% of direct mode.
Chris

Headstart 6i - no client/server support?

Hello,
I have received today list of some new features of Headstart 6i (Beta supposed to be available this week).
In the new features list, here what it is wrote:
<<Because of known issues associated with running Forms 6i in a client-server environment, Headstart 6i will not be certified for client-server deployment. Applications can still be deployed client-server, but Headstart will not attempt to identify and overcome client-server specific issues.>>
Is someone have heard about these "known issues" in client/server?
Thank you
Jean
null

The problems are GUI issues.
The basic problem is this:
Headstart can either be optimized to display in Webforms with the Oracle Look and Feel, or it can be optimized to display in client-server mode. You can't do both at the same time. Since we had to pick one or the other, we chose the Webforms OLAF GUI settings.
You can customize the template package to revert to the client-server GUI settings if you like. This basically involves changes to the visual attributes in the object library, removing the use of the keyword 'automatic' and setting actual colors for foreground color, background color and fill pattern.
There are also problems with displaying the menu and Smartbar. You should remove the icons that are displayed in the menu. If an entire menu group is disabled, the associated icons on the Smartbar disappear instead of greying out (no workaround).
There might also be other GUI issues. As stated, we have not tried to identify or resolve these issues.
Regards,
Lauri

Headstart Designer 6i. Install into Production Environment as Client/Server

Hi,
We are using designer 6i with headstart installed. We already have it installed in one 8i database running on an NT server but now we need to deploy it in a different 8i database. The associated documentation relates to a web server enviroment. Can you provide the steps or documentation to install this on a client server environment?
Thanks

Paul,
You're right.. You shouldn't ;-D But a good hack is hard to resist. Therefore, create the following view:
create or replace view my$version as select replace(banner,'Personal ') banner from sys.v_$version;
Now, create a synonym named v$version to this view.
The neatest way to do this is to create the view my$version view in your HST schema, and create a grant and a private synonym to this view in the schema's that you use to run the Forms:
create synonym v$version for <HST_OWNER>.my$version
But if you're bold enough you could also remove the existing public synonym v$version (which points to sys.v_$version), install the my$version view in the SYS schema and create a new public synonym:
create public synonym v$version for my$version;
And remember, you never heard this from me ;-D
Kind regards,
Peter
null

VPN Site-to-Site or VPN Client Server with Cisco IP Phone 8941 and 8945

Hi everyone,
I decide to deploy a CUCM (BE6K platform), SX20, and IP Phone 8941/8945 on Head Office and Cisco SX10 and IP Phone 8941/8945 for branch offices (actually 9 branch offices).
The connection will use internet connection for HO and each branch offices.
And the IT guy want to use kind a VPN client server or VPN site-to-site for the connection through internet,
what kind of VPN client server or VPN site-to-site that recommended for this deployment?
and what type of Cisco router that support that kind of VPN (the cheapest one will be great)?
So the SX10 and IP Phone 8941/8945 in branch offices can work properly through internet connection?
please advise
Regards,
Ovindo

Hi Leo,
technically, the ipsec users will not use up any premium license seats, so if you have 10 ipsec users connecting first, the premium seats are still free and so you can then still have 10 phones/anyconnect users connect.
However, the 250 you mention is the global platform limit, so it refers to the sum of premium and non-premium connections. Or in other words, you can have 240 ipsec users and 10 phones, but not 250 ipsec users and 10 phones.
If 250 ipsec users and 10 phones would try to connect, it would be first-in, first-served, e.g. you could have 248 ipsec users and 2 phones connected.
Note: since you have Essentials disabled I'm assuming you are referring to the legacy "Cisco vpnclient" (IKEv1 client) which does not require any license on the ASA. But for the benefit of others reading this thread: if you do have Anyconnect clients (using SSL or IPsec/IKEv2) for which you currently have an Essentials license, then note that the Essentials and Premium license cannot co-exist. So for e.g. 240 Anyconnect users and no phones, you can use Essentials. For 240 Anyconnect users and 10 phones, you need a 250-seat Premium license (and a vpn phone license).
hth
Herbert

Try to use one comupter as both server and client

Hello, everyone, I am just trying to use my own computer as both server and client to test some codes about networking. For example, use the sample code in java tutorial which is used to test Echo server(code is listed below). Is there anything I have to do to set my computer, such as set my hostname or something like that?
I am a pure newbie. And the purpose of this question is to test some code including socket on one PC without connect to internet.
I have tried to change the name "taranis" in the following code to the computer name of my own PC, but it doesn't work, and said: Couldn't get I/O for the connection to: (my computer name).
import java.io.*;
import java.net.*;
public class EchoClient {
public static void main(String[] args) throws IOException {
Socket echoSocket = null;
PrintWriter out = null;
BufferedReader in = null;
try {
echoSocket = new Socket("taranis", 7);
out = new PrintWriter(echoSocket.getOutputStream(), true);
in = new BufferedReader(new InputStreamReader(
echoSocket.getInputStream()));
} catch (UnknownHostException e) {
System.err.println("Don't know about host: taranis.");
System.exit(1);
} catch (IOException e) {
System.err.println("Couldn't get I/O for "
+ "the connection to: taranis.");
System.exit(1);
     BufferedReader stdIn = new BufferedReader(
new InputStreamReader(System.in));
     String userInput;
     while ((userInput = stdIn.readLine()) != null) {
     out.println(userInput);
     System.out.println("echo: " + in.readLine());
     out.close();
     in.close();
     stdIn.close();
     echoSocket.close();

Did you write the EchoServer and start it on your
machine, listening on port 7?
You can have the client and server running on the same
machine or different machines, but they have to be
separate pieces of software.
Write a separate EchoServer class that starts up and
listens on that port. Then start the EchoClient and
make the connection.
%yeah, I didn't wrote the EchoServer class. But I thought it is automaticly included and therefore has run once I start my computer.
If I write a EchoServer class, then how should I set the host name of the EchoClient, just simply change "taranis" to my computer name (change "echoSocket = new Socket("taranis", 7);" to echoSocket = new Socket("(my comptuer name)", 7);"?

Can i use same Server for server side and client??

Hi,
i m developing webservices in java and using two different server for server side and client.
e.g. i m using one tomcat server on a machine to run webservice and again using one more tomcat server on client side at different machine.
and hence it need two tomcat server.
But i want only one server to run webservice and client.
So please help me out...
Thanks

Hi,
It is always recommended to maintain different servers
REgards,
Ravi.

Looking for a client/server that supports multiple protocol and delivery

Hi all, I don't know if this the right place to ask my question,here it goes.
I am looking to develop a client-server that supports multiple protocols such as HTTP, HTTPS etc. I am looking for a framework( i don't know if that is correct or I need some kind of web-service (soap etc)) that would manage connection, security etc. I would like to like to devote most of my time in developing business objects with multiple delivery mechanism such as sending serilized java objects, xml message or soap message, or in some case JMS message as well. So I need a client server that can come in via TCP/IP or HTTP or anyother industry standard protocol and I should be able to service him with pub/sub model and also request/response model as well.
I don't know if I had explained what I need, I would like to know what technologies I should be evaluating and which direction I should be heading... Also the server I'm developing should be free of Java constraints if needed...
Also this service is not webbased service as now but if need arises I should have a flexibilty to make them web enabled in future. Also I would like to work with open source webservers or appservers if I need

Inxsible wrote:I installed i3 - along with the i3status - which I still have to figure out. I am liking what I see as of now. It reminds me of wmii -- when I used it way back when. However I do not like the title bar. I would much rather prefer a 1 px border around the focused window.
"i3 was created because wmii, our favorite window manager at the time, didn't provide some features we wanted (multi-monitor done right, for example), had some bugs, didn't progress since quite some time and wasn't easy to hack at all (source code comments/documentation completely lacking). Still, we think the wmii developers and contributors did a great job. Thank you for inspiring us to create i3. "
To change the border of the current client, you can use bn to use the normal border (including window title), bp to use a 1-pixel border (no window title) and bb to make the client borderless. There is also bt which will toggle the different border styles.
Examples:
bindsym Mod1+t bn
bindsym Mod1+y bp
bindsym Mod1+u bb
or put in your config file
new_window bb
from: http://i3.zekjur.net/docs/userguide.html (you probably already found that by now )

JAX-RPC: Web service where both server /and/ client are services?

I'm considering a two-way Web service in an application that I'm designing and would appreciate any insight from prior experience.
This will be a standard client-server application, where the server broadcasts messages and the client, in turn, can request further information about the messages that it receives (think RSS-style updates, but with more interaction). [strong]However[strong], instead of having the client poll the server for new messages, I'd like to have the client register itself with the server as a message listener. The server is currently implemented as a Web service on port X. I'd like the client to register itself with the server, and then receive messages on port Y.
I'm considering a design like the following (note the directions of the inheritance arrows):
( Server class ) --<-- ( RemoteServerImpl class ) -->-- / RemoteServer interface /
|
( Listener class ) --<-- ( RemoteListenerImpl class ) -->-- / RemoteListener interface /Some considerations:
* In both cases, the client and server are broken out into an interface and two classes. The non-Remote classes implement all of the functionality. The RemoteX interfaces declare the methods that will be exposed via RPC, and the RemoteXImpl classes are empty subclasses of the non-Remote classes.
* In order to register itself with the Server, the Listener must be passed in as a method parameter, and thus conform to the JAX-RPC specification for value types. However, value types can't implement java.rmi.Remote.
* Only the Server and Listener know anything about each other.
* This application can be run either remotely or locally, depending on how the Server and Listener are instantiated. For the latter, they're instantiated directly and the listener is registered with the Server. For the latter, the RemoteX interfaces will point to the service stubs.
* I don't see an easy way to allow both the Server and the Listener to have references to each other (at least now without some casting magic), since one of them will have to conform to the value type specification, which requires that all of its fields must be valid JAX-RPC types.
I've got the Server portions implemented and working, and am now going to turn to the Listener stuff. Can anybody see a fundamental problem with this design, or can suggest something better?

Hi There,
Because web services are fundamentally a Services-oriented architecture, and not an Object-oriented one, you are unable to pass objects by reference (only by value) hence the design decision to not allow rmi.Remote objects to be sent across the wire. The best way would be to abstract an object which does itself not represent a server, but a route to a server (lets call it ServerPointer). This must contain address, port, service info, etc.
Now, in your Server class, create a factory method which takes a ServerPointer, and returns a reference to a
remote server, e.g. protected Server createServer( ServerPointer p ) (using the necessary APIs to get such a reference)
Hope this helps...

Headstart and Client/Server (C/S)

Similar Messages

Maybe you are looking for