At 04:19 PM 07/14/2000 +0100, you wrote:
>Anyone had any success with the above?
>
>i just downloaded the jdk1.3beta that was released yesterday, and after
>fiddling around with classpaths etc i finally got jboss running and
>autodeploying w/o errors.
>
>problem is from the client side, when i try to access a bean, i always end
>up with something like:
>javax.naming.CommunicationException [Root exception is
>java.io.StreamCorruptedException: Type code out of range, is 125]
>
>this happens both with my own beans and the supplied Bank example beans,
>already when doing a lookup.
>btw, whose idea was it to bundle an example bean jar, with no
>documentation, no source and no client/test apps?!? or maybe it's just me
>who can't find them.
>
>it should be said that i've never managed to get any beans working on
>jboss, but the ones i've tried have all been running successfully on other
>ejb servers (orcas4.0.0)
>
>cheers for what looks like a great project, though maybe you should try to
>catch up on documentation some time (Rikard - dokumentera mera!).
I've added some more specifics on how to do this to the Bugzilla bug we're
using to keep track of changes to the jBoss documentation, bug #133. Here's
what I said....
Comments on jBoss Installation Documentation
1) The first heading, "What this article is about", is not needed.
2) In the first paragraph, the first mention of jBoss should have a link to
the jBoss Web site, <A HREF="http://jboss.org/">jBoss</A>. At the end of
the first paragraph, there shouldn't be an apostrophe in "one's I've assumed".
3) In the second paragraph (beginning with "jBoss is an implementation of
the EJB 1.1 specification,") the phrase "EJB 1.1 specification," could be a
link to the spec, <A HREF="http://java.sun.com/products/ejb/docs.html">EJB
1.1 specification,</A>
4) In the sentence, "In this it is similar to Sun's `J2SDK Enterprise
Edition' (J2EE), but jBoss is much more single-minded than J2EE", would be
a good place to explain the chief advantages of jBoss over the Sun J2EE
Reference Implementation. Instead of "but jBoss is much more single-minded
than J2EE" it would be better to say, "but jBoss is an Open Source
implementation which can be used for commercial purposes" (or whatever you
think is the chief advantage of jBoss over J2EE). So here's how I think the
sentence should be worded: "It is similar to Sun's `J2SDK Enterprise
Edition' (J2EE), but jBoss is an Open Source implementation which can be
used for commercial purposes."
5) In the next paragraph, the sentence "You get no support, of course" is
misleading to people who are not familiar with Open Source projects.
Support is always available for Open Source projects, but not directly from
the vendor (because there are many cooperating "vendors"). This paragraph
actually deals with two completely different topics (hot deployment and
Open Source/GPL/support). I think it should be broken into two paragraphs,
and the new paragraph dealing with Open Source/GPL/support should be:
'jBoss is distributed under the <A
HREF="http://www.gnu.org/copyleft/copyleft.html">GNU public licence</A>,
which means that it's free, even for commercial work, and is likely to
remain that way. As with most Open Source projects, support is available
from third-party vendors and an active <A
HREF="http://jboss.org/mailing.htm">mailing list</A> of enthusiastic
volunteers.'
6) The paragraph starting with "The main weakness of jBoss is its
documentation" is the opposite of a self-fulfilling prophecy: if this
article (and its cousins) do their job, the weakness goes away. It might be
more accurate to say, "Like jBoss itself, the jBoss documentation is still
evolving. You can start with this article, describing step-by-step how a
simple EJB can be created, deployed and tested on the jBoss server, then
use the mailing list and other documentation (to be added later) for more
advanced topics."
7) On the next page, <http://www.web-tomorrow.com/jboss1.html>, the first
paragraph is a bit confused. "The first step" isn't really the first step
-- the rest of the paragraph says the very first step is to get your JDK
running, then it doesn't give detailed directions on how to do that. It
seems to me that we need details, but that they should be in a separate
page, off of the main flow of the installation instructions. (After all, if
the reader already has JDK 1.3 installed, there's no need to go through the
details.)
First, login to your Linux system as root. You must have administrator
access to install the JDK correctly.
Download JDK 1.3. Here are two sources: Sun's <A
HREF="http://java.sun.com/j2se/">Java<SUP><FONT SIZE="-2">TM</FONT></SUP> 2
SDK, Standard Edition, v 1.3</A> and IBM's <A
HREF="http://www.ibm.com/java/jdk/download/index.html">JDK</A>. (We'll give
step-by-step instructions for Sun's version.)
You can use your favorite Web browser to download the JDK to your Linux
machine. If you use Lynx, the text-only browser that comes with Redhat, you
can download it directly to your Linux machine, even if your Linux machine
doesn't have a graphical browser installed. Use this command:
# lynx http://java.sun.com/j2se/1.3/
Then navigate the page, using the down arrow until you highlight the link
to download the Linux version of the Java(TM) 2 Platform, Standard Edition
version 1.3 (currently, the link has the text "Linux Beta" but that should
change soon). Press the enter key. The next page asks you to register or
login. Use the down arrow to move between fields, type your user ID and
password, then use down arrow to "Log in" and press enter. Accept any
cookies the server throws at you.
On the next page (titled "Java(TM) 2 SDK, Standard Edition, v 1.3.0 Beta"),
use the arrow keys to navigate to "One large bundle" then press enter on
"Select Install Format." If you're using Redhat Linux, you can arrow down
to "Red Hat (RPM)" and press enter. Arrow down to "continue" and press
enter again. (If you're not using Redhat, choose gunzip-tar.) This will
bring you to the "License & Export for Java(TM) 2 SDK, Standard Edition
1.3.0" page.
On the "License & Export for Java(TM) 2 SDK, Standard Edition 1.3.0" page,
read through the license and arrow down to "ACCEPT" then press enter.
On the next page, "Download Java(TM) 2 SDK, Standard Edition 1.3.0", arrow
down to "FTP download" and press enter. (Or, if you're behind a firewall,
arrow down to "HTTP download" and press enter.) Either way, Lynx will ask
you if you want to D)ownload, or C)ancel. Press D. Now you wait for a
while. After Lynx eventually downloads the 39 megabyte RPM file to a
temporary location, it will ask you for a file name. Use the default name,
"j2sdk-1_3_0-beta-linux.rpm". When Lynx is done, you can quit by pressing Q.
Now under Redhat, you can type this command:
# rpm --install j2sdk-1_3_0-beta-linux.rpm
This will install the JDK in this directory:
/usr/java/jdk1.3
Remember that directory. It will be important later on.
If you have installed a different JDK (like the one from IBM) or you have
installed it in a different directory, you'll need to know the name of that
directory. You can find it like this:
# find / -name javac -print
This will print out a list of all of the directories containing the Java
compiler, javac. My system still has the older JDK 1.2.2 and the IBM JDK on
it, so my system tells gives me this result:
/usr/local/jdk1.2.2/bin/i386/native_threads/javac
/usr/local/jdk1.2.2/bin/i386/green_threads/javac
/usr/local/jdk1.2.2/bin/javac
/usr/java/jdk1.3/bin/i386/green_threads/javac
/usr/java/jdk1.3/bin/i386/native_threads/javac
/usr/java/jdk1.3/bin/javac
find: /proc/6/fd: Permission denied
/opt/IBMJava2-13/bin/exe/javac
/opt/IBMJava2-13/bin/javac
From that list, I can tell that my newly-installed Sun JDK 1.3 is in
/usr/java/jdk1.3. (I can also see that I don't have a floppy in the drive
-- that's what the "Permission denied" message is about.)
You now need to add /usr/java/jdk1.3/bin directory to your PATH environment
variable. You use one of the following commands depending on your shell.
.bash:
PATH=$PATH:/usr/java/jdk1.3/bin
csh, tcsh:
# set path=($PATH /usr/java/jdk1.3/bin)
sh:
# PATH=($PATH /usr/java/jdk1.3/bin); export $PATH
You should also add these lines to your .bash_profile, .profile and .cshrc
files so you won't have to do this every time you login. Note that your
shell will usually find the first occurrence of a program along your path,
so if you have more than one version of Java installed, you should add only
one to your path at any time. To find out the path to the version of Java
that will run when you enter a java command, type this:
# which java
Adjust your path until this command tells you /usr/java/jdk1.3/bin. You may
need to edit your .bash_profile (or other file) then do this command:
# source .bash_profile
(Meanwhile, back at jBoss-specific installation procedures)
8) In the paragraph starting "It doesn't matter very much where you install
jBoss" I don't happen to agree with your suggestion of /usr/lib/jboss/. I
think local applications belong in /usr/local/, so I like
/usr/local/jboss/. (What does everyone else think? Where is jBoss installed
on your Linux system?)
9) The paragraph starting 'jBoss is distributed as a ZIP archive,' could be
modified to include a link to the jBoss binary like this: 'jBoss is
distributed as a <A HREF="http://www.telkel.com/binary/jboss.zip">ZIP
archive,</A>'
10) Before telling the reader how to do the jar/unzip command, we ought to
suggest how to create the directory and download the Zip file to the Linux
machine, something like this:
mkdir /usr/local/jboss
cd /usr/local/jboss
lynx -source http://www.telkel.com/binary/jboss.zip > jBoss-2.0.zip
jar xvf jBoss-2.0.zip
rm -f jBoss-2.0.zip
cd bin
java -jar run.jar
(That kinda takes all of the guessing out of the process, doesn't it? All
the reader needs to do is to copy/paste that block of commands to the Linux
prompt and jBoss is up and running. We could label this the "Quick Start
Procedure.")
11) In file <http://www.jboss.org/jboss1.html>, in the paragraph starting,
"If the JDK binaries aren't in the PATH" won't the previous command (java
-jar run.jar) fail if the JDK binaries aren't in the PATH? It seems like a
good idea to organize these "Things to watch out for" a little better, like
a list of error messages often seen during start-up and a corrective action
for each, something like this:
<TABLE BORDER=1 CELLSPACING=0 CELLPADDING=5 BGCOLOR="#FFFF80">
<TR VALIGN="TOP">
<TH>Error Message</TH>
<TH>Corrective Action</TH>
</TR>
<TR VALIGN="TOP">
<TD>bash: java: command not found</TD>
<TD>The JDK binaries aren't in the PATH. Check your PATH environment
variable and try the command<PRE># which java</PRE>See the section on
downloading and installing the JDK.</TD>
</TR>
<TR VALIGN="TOP">
<TD>[hypersonic] System.run/init: database is in use by a different
process</TD>
<TD>The JDK binaries aren't in the PATH. If you get this message, the
SQL server hasn't started, and there will be a whole heap of other errors
later. Check your PATH environment variable and try the command<PRE># which
java</PRE>See <A HREF="#somewhere_else">the section on downloading and
installing the JDK.</A></TD>
</TR>
<TR VALIGN="TOP">
<TD>Failed to load Main-Class manifest attribute from<BR>
run.jar</TD>
<TD>You're not in the right directory and Java can't find file
run.jar. Try this:<PRE># cd /usr/local/jboss/bin</PRE></TD>
</TR>
<TR VALIGN="TOP">
<TD>[Webserver] java.net.UnknownHostException:...</TD>
<TD>The jBoss server tries to establish its host's IP number when it
starts up. If it can't do this, you will see this error message. This
problem is quite common on systems that get their network configuration
parameters dynamically. For example, if your computer connects to the
Internet by a dial-up connection, and you cancel the dial-up process, you
may end up with a hostname that doesn't make sense. You can re-establish
your connection or you can temporarily reset your hostname to `localhost'
like this: <PRE># hostname localhost</PRE></TD>
</TR>
<TR VALIGN="TOP">
<TD>[Webserver] java.net.BindException: Address already in use</TD>
<TD>This probably means you already have jBoss running and you're
trying to start it a second time. Use this command:<PRE># ps ax | grep
run.jar</PRE>to determine the process ID of the previous attempt to invoke
jBoss then use kill -9 (pid) to kill that PID. Restart jBoss.</TD>
</TR>
<TR VALIGN="TOP">
<TD>[Container factory] No configuration chosen. Using default
configuration</TD>
<TD>No corrective action necessary. This means that a particular EJB
did not have a jboss.xml file in its META-INF directory (which is inside
the EJB's .jar file in the deploy/ directory). This isn't a real problem;
jBoss will just use its default configuration. If you want to change the
configuration, see <A HREF="#somewhere_else">the section on jboss.xml.</A></TD>
</TR>
<TR VALIGN="TOP">
<TD>[Container factory] javax.management.InstanceNotFoundException:
DefaultDomain: service=Webserver</TD>
<TD>TBD</TD>
</TR>
<TR VALIGN="TOP">
<TD>[JMX RMI Adaptor] Started<BR>
java.net.BindException: Address already in use</TD>
<TD>TBD</TD>
</TR>
</TABLE>
12) I think it would be worthwhile to tell readers how to set up jBoss in
/etc/rc.d/init.d/ so that it will restart every time the system reboots.
13) Per the e-mail note listed as another attachment to bug #133, I think
it would be worthwhile to give details on how to configure jBoss with MS
SQL Server. (This may be beyond the scope of the introductory installation
and how-to guide, and may be put in an "Advanced jBoss Installation" section.)
14) It would also be worthwhile to give details on how to configure jBoss
with MySQL. ("Advanced jBoss Installation" section.)
15) We should have a parallel documentation set detailing installation on
Windows NT and any other commonly-used platforms. One way to implement this
would be to use servlets or JavaScript and ask the reader which system he's
using, then customize the documentation for that system.
-- Ken Jenks, http://abiblion.com/
Tools for reading.
--
--------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Problems?: [EMAIL PROTECTED]
