sub venkat

From: Henrik Sjostrand
Subject: Re: Using ANT for production deployment
Newsgroups: ibm.software.websphere.application-server
Date: 2004-06-04 11:40:43 PST


subkat wrote in message news:<91e6fa21.0406031150.60e9c27@posting.google.com>...
> Presently, our production deployment is done using a jacl script.
> Apart from deployment, the jacl script also has functions to start
> cluster(s), stop cluster(s), restart cluster(s), check cluster status,
> check server status, stop app, start app, check app status etc.
>
> My company had me researching on using ANT. What would be compelling
> reasons to use ANT in place of the existing jacl script? I have
> researched on the web and found some lukewarm arguments.

I'd say stick with your existing, working, JACL scripts and use ANT
for your Java build process. Sure you can use ANT also for controlling
WebSphere, but I prefer to use JACL (wscp/wsadmin) for this.

ANT is not a scripting language, it's a declarative language which is
very suitable for the build process. But in the deployment process I
see the advantages of using a scripting language with things like
conditional statements etc.

Some people will argue and say that you can use ANT for both build and
deployment and, of course, you can. And ANT can also handle
conditional statements but not as easy as JACL. One advantage of using
ANT instead of JACL, however, would be that you only have one tool and
syntax to learn to control both processes.

/Henrik Sj�strand

Forgot to mention we use WebSphere 5

Presently, our production deployment is done using a jacl script. Apart from deployment, the jacl script also has functions to start cluster(s), stop cluster(s), restart cluster(s), check cluster status, check server status, stop app, start app, check app status etc.

My company had me researching on using ANT. What would be compelling reasons to use ANT in place of the existing jacl script? I have researched on the web and found some lukewarm arguments.

Presently, our production deployment is done using a jacl script. Apart from deployment, the jacl script also has functions to start cluster(s), stop cluster(s), restart cluster(s), check cluster status, check server status, stop app, start app, check app status etc.

My company had me researching on using ANT. What would be compelling reasons to use ANT in place of the existing jacl script? I have researched on the web and found some lukewarm arguments.

I did not include those green-alien smileys. I have no idea how they creeped in!!

Sub Venkat

The ANT script below tries to stop an app named "testapp". I am running it on the host machine via "ws_ant". IT BOMBS. Any insight is really really appreciated..

____________________________________________________
Buildfile: /tmp/ant_test/stopapp.xml
____________________________________________________
<?xml version="1.0"?>
<project name="test_deploy2" default = "stopapp" basedir=".">



<target name="stopapp" >
<taskdef name="wsStopApplication" classname="com.ibm.websphere.ant.tasks.StopApplication"/>

<wsStopApplication
wasHome="a/b/c/AppServer"
application="testApp"
conntype="SOAP"/>


</target>

</project>
____________________________________________________

stopapp:

[wsStopApplication] Stopping Application [testApp]...

[wsadmin] WASX7209I: Connected to process "dmgr" on node xYZManager using SOAP connector; The type of process is: Deploym
entManager
[wsadmin] WASX7017E: Exception received while running file "/tmp/wsant57218jacl"; exception information: com.ibm.bsf.BSFExc
eption: error while eval'ing Jacl expression: com.ibm.ws.scripting.ScriptingException: WASX7025E: Error found in String "WebS
phere latform=common,cell=Lab,version=5.0,name=ApplicationManager,mbeanIdentifier=ApplicationManager,type=ApplicationM
anager,node=xYZ,process=GoodApp_TST_JVM_0
[wsadmin] WebSphere latform=common,cell=Lab,version=5.0,name=ApplicationManager,mbeanIdentifier=ApplicationManager,t
ype=ApplicationManager,node=xYZ,process=testapp_JVM_0
[wsadmin] WebSphere latform=common,cell=Lab,version=5.0,name=ApplicationManager,mbeanIdentifier=ApplicationManager,t
ype=ApplicationManager,node=xYZ,process=server1
[wsadmin] WebSphere latform=common,cell=Lab,version=5.0,name=ApplicationManager,mbeanIdentifier=ApplicationManager,t
ype=ApplicationManager,node=xYZManager,process=dmgr"; cannot create ObjectName.
[wsadmin] Java Result: 105

[wsStopApplication] Stopped Application [testApp]
____________________________________________________


thanks
Sub Venkat

I am sorry, I did not read your message properly earlier. The PVCS here is version 7.5

Hi,

Trying to write an ant script for (re)deploying an ear (that is on my local XP box) onto aix server (ND). Now this involves stoppping app--> uninstalling app--> installing (newer version) app--> starting the app.

The app is clustered across two nodes. Do I have to do the above process for each node separately or is there a way to do it at the cluster level?

I am using the wsStopApp, wsUninstallApp, wsInstallApp, wsStartApp tasks of ANT.

thanks

At the time of posting this message, I was using ANt 1.4.1. But later I shunned that and downloaded a binary version 1.6.1. from apache, and am NOT running it in WSAD but from command line. It works with the PVCS task also.


thanks.

Hey about deleting the directory (obviously along with its contents), then making a directory with the same name? (<mkdir> task!)

Thanks Paul,

My JAVA_HOME points to a JRE, not JDK. So I downloaded a JDK and reset the JAVA_HOME variable.

thanks.

Hi,
I just installed ANT on Win XP box, and after setting up the ANT_HOME and the ant/bin in the path,
ran "ant -version". Th eoutput follows:
______________________________________________________
Unable to locate tools.jar. Expected to find it in c:\progra~1\java\lib\tools.jar
Apache Ant version 1.6.1 compiled on February 12 2004
_______________________________________________________________

Where can I find tools.jar? I could'nt find it in the ANT distribution.

thanks,
Sub cat

HI,
I have ANT 1.4.1 running in WSAD. When I try to use the "PVCS" task in a build file, the script bombs with the error : "Build exception: Failed executing: pcli lvf -z -aw -prS:\Wsadproj\WSWSAD~1 /myproj"

Two of several thoughts that I have are:
A. If it went that far as to say "pcli" command doesnt work, then may be it is supported.
B. May be I did'nt properly set the PCLI_ID (the repository is remote, but I do have access). I did set it from command line.

thanks in advance for your time
sub cat

I know my message earlier (tiltled same as this) was verbose. Please let me know if I need to rephrase stuff or any thinng. But I need help bad.

thanks,
sub cat

I try to run the ANT pvcs task from within WSAD (Right click on build.xml-->"Run Ant").
1. Now this build.xml file, I placed in the project workspace (of WSAD).
2. Before I placed the build.xml file, I imported the project into this work space in WSAD manually from the PVCS repository,

then copied the build.xml file into the project folder (which is within the WorkSpace in WSAD).
3. That is how I am able to access the ant script (build.xml) from within WSAD (ie., since I placed the build.xml file in the

same workspace as the project).
4. This is the ANT task I m trying to run:

<pvcs repository="S:\Wsadproj\WSWSAD~1" pvcsproject="/myproj"/>

What I am expecting this to do is get the latest project copy of "myproj" from the remote repository on network-mapped S

drive, into my WSAD workspace from where I am running the ANT script (build.xml).
5. Now the question is: the ant script throws this exception:
Build exception: Failed executing: pcli lvf -z -aw -prS:\Wsadproj\WSWSAD~1 /myproj
6. I do have access to the repository since I was able to run the command:
pcli lvf -z -aw -prS:\Wsadproj\WSWSAD~1 /myproj
after setting "PCLI_ID" from a command window.

Any help is greatly appreciated.

thank you.