[JBoss-user] [Clustering/JBoss] - Re: Sub-partitions, is it in code now or not?
slaboure wrote : well, it is easier: just subscribe to jboss-user and you will get each post by e-mail :) There's a better way if you don't want to get every forum. When you log on go to the 'Watched Forums' tag and choose the ones you want, no need to subscribe to jboss-user. Kev View the original post : http://www.jboss.org/index.html?module=bbop=viewtopicp=3823681#3823681 Reply to the post : http://www.jboss.org/index.html?module=bbop=postingmode=replyp=3823681 --- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356alloc_id=3438op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence CMP/JBoss] - Re: callbacks from CMP or caching layer
Don't expose the CMP field in the interface, instead expose a method that sets the CMP field and your flag. View the original post : http://www.jboss.org/index.html?module=bbop=viewtopicp=3823276#3823276 Reply to the post : http://www.jboss.org/index.html?module=bbop=postingmode=replyp=3823276 --- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356alloc_id=3438op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [Persistence CMP/JBoss] - Re: callbacks from CMP or caching layer
Forgot to add that you also need to set your flag in the ejbCreates, the ejbLoad and the ejbActivate. View the original post : http://www.jboss.org/index.html?module=bbop=viewtopicp=3823278#3823278 Reply to the post : http://www.jboss.org/index.html?module=bbop=postingmode=replyp=3823278 --- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356alloc_id=3438op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] [EJB/JBoss] - Re: ClassCastException
The purpose of the PortableRemoteObject.narrow method is to handle the casting when the object uses IIOP as the transport mechanism (corba or RMI/IIOP) and has nothing to do with your problem. You should use this whenever downcast an object that may be using IIOP, the lookup of a home interface being one such occassion. The fact that you can get it to work when you set the thread context loader suggests that you have multiple copies of that interface in different class loaders. The cast will only work if you are casting to a class/interface loaded with the same classloader. Add some debug before the cast to walk the class and interface hierarchy of the objects, printing out the class AND classloader in each case. My guess is that you will find your FooHome interface loaded by multiple classloaders, perhaps in multiple jars. Identify and remove the duplicate entries and your problem will be solved. :-) View the original post : http://www.jboss.org/index.html?module=bbop=viewtopicp=3822654#3822654 Reply to the post : http://www.jboss.org/index.html?module=bbop=postingmode=replyp=3822654 --- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356alloc_id=3438op=click ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
Re: [JBoss-user] Optimistic locking and Commit Option A
Pete Beck wrote: Yep, we've walked that same path too. Unfortunately there are few long running transactions which need write access and these seem to screw things up. We also have long running transactions, calling potentially external applications from our workflow app, but we have chosen to restructure the app and follow the readers/writers strategy imposed by the pessimistic/read-only combo. We defer most of our writes as late as possible to cut down on the contention and use separate transactions for those we can't. By the way, does anyone know how wildcarded read-only methods (e.g. get*) apply when the method has been inherited from another class? It applies to all matching members of the defined interface used to expose it. In other words if I have getSomething() in class A, and then create class B which inherits from class A. Will a get* rule in class B also apply to the getSomething() method inherited from class A? Good luck. Kev --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBoss-net does not like calling EJB method with transaction type = never
If the bean is 'never' it won't work anyway (because the calls to the SB with 'Required' TX attribute is done withiout a TX context). RequiresNew should be used in that case. Are you not talking about mandatory? Required will create a tx context if one does not exist. Kev --- This SF. Net email is sponsored by: GoToMyPC GoToMyPC is the fast, easy and secure way to access your computer from any Web browser or wireless device. Click here to Try it Free! https://www.gotomypc.com/tr/OSDN/AW/Q4_2003/t/g22lp?Target=mm/g22lp.tmpl ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Commit Option Confusion
Hiya Bill. I found that the database we use is expensive when it comes to preparing statements and it has bugs when prepared statements are reused (why else would you use prepared statements ;-)). You mean you created one big SQL call with JOINs and shit? If so, would you be interested in contributing this code? Sorry, that's not what I meant. The SQL generated is still the same as it was originally. The biggest impact was the load group strategy (or read ahead strategy I guess). Our code makes a lot of use of relationships and the generation of certain queries were not being handled as efficiently as we would like. The JDBCLoadRelationCommand was breaking up the lists based upon the PK of the parent entity. I wanted the parent groupings to be inherited by the relationship so that it could batch load the entities. A simple example is two beans related by a CMR. If you executed a finder on the first bean and retrieved, say, 10 PKs, then these would be all be loaded in a single query. If you then executed the load relationship command then it would group the child beans based on the parent PK. Loading the child beans would then result in 10 batch requests, one based on each parent, instead of the single one that is perhaps necessary. The change I made is to allow certain relationships to be grouped based on the grouping containing the parent PK. This would result in two accesses (one for each bean type) instead of eleven for the above scenario. I also changed the ReadAheadCache code so that it is now a two stage cache, local tx and global, so that the cache can propogate across the tx boundary. There's other stuff in there to prevent rollback info being cached and entity bean info from being overwritten. I think that is a good idea. I was able to combine this with the previous strategy because we knew that certain queries from our front end are usually followed by accesses for the same set of beans. Getting this information into the cache and keeping it there made a great improvement. The other performance change that I made, although its impact is not as great as the strategy, is to cache the dynamic query information (we use it a bit in two of our beans). I put a LRU cache into JDBCDynamicQLQuery so that it could keep track of the generated SQL, the selected object (or field) and the input parameters without requiring a recompilation. Kev This e-mail and any files transmitted are confidential and is solely for the intended recipient. Unauthorised use, dissemination, distribution, copying, reproduction, modification or publication is strictly prohibited. If you have received this e-mail in error, please notify the sender and permanently delete the e-mail from your system. Opinions expressed in this e-mail are those of the individual and not necessarily those of Orchard Information Systems Ltd. It is the responsibility of the recipient to ensure this e-mail is free from viruses. Orchard Information Systems Ltd accept no liability for any damage caused by any virus transmitted by this e-mail. Incoming and outgoing e-mail messages are routinely monitored for compliance with Orchard Information Systems Ltd's Internet Usage Policy. --- This SF.Net email sponsored by: ApacheCon 2003, 16-19 November in Las Vegas. Learn firsthand the latest developments in Apache, PHP, Perl, XML, Java, MySQL, WebDAV, and more! http://www.apachecon.com/ ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Commit Option Confusion
I am missing something or is my understanding of Commit options flawed? I moved our app from B to A, added a new load group strategy and it flies. It achieved a 90% speedup in our app. The best way to see what is happening is to enable the cmp debugs and the ReadAheadCache info (yes we're still on 3.0.2) and try to understand what is happening. If you can get the information into jboss ASAP then it can really fly. Kev --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Commit Option Confusion
load group actually worked to improve your performance? Do you have BLOBs or something? I don't see this with my specj benches. I found that the database we use is expensive when it comes to preparing statements and it has bugs when prepared statements are reused (why else would you use prepared statements ;-)). I made two changes to offset this. I created a new load group strategy, called it deep-load, that propogated the load group pk information through the relationships (we're still loading all fields). I also changed the ReadAheadCache code so that it is now a two stage cache, local tx and global, so that the cache can propogate across the tx boundary. There's other stuff in there to prevent rollback info being cached and entity bean info from being overwritten. Just using the first had a big impact on our code as we follow relationships heavily. The number of prepared statements we used dropped significantly. The second just meant we weren't throwing the data away after the transaction finished. Transactional consistency was my biggest concern with the second change. Our tests haven't shown up any problems so far but they may still exist. The mitigating factor is that we are just using a single instance of jboss with commit A. The code would certainly break under other circumstances. Kev This e-mail and any files transmitted are confidential and is solely for the intended recipient. Unauthorised use, dissemination, distribution, copying, reproduction, modification or publication is strictly prohibited. If you have received this e-mail in error, please notify the sender and permanently delete the e-mail from your system. Opinions expressed in this e-mail are those of the individual and not necessarily those of Orchard Information Systems Ltd. It is the responsibility of the recipient to ensure this e-mail is free from viruses. Orchard Information Systems Ltd accept no liability for any damage caused by any virus transmitted by this e-mail. Incoming and outgoing e-mail messages are routinely monitored for compliance with Orchard Information Systems Ltd's Internet Usage Policy. --- This SF.Net email sponsored by: ApacheCon 2003, 16-19 November in Las Vegas. Learn firsthand the latest developments in Apache, PHP, Perl, XML, Java, MySQL, WebDAV, and more! http://www.apachecon.com/ ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Dynamic Addition of Roles using JAAS
I meant here authentication. But I have some doubts - is it done in the end of successful authentication, or just before the first authorization? - I don't know it. The information is cached after a successful authentication and expires after a specified period. The login module is called if the user is not in the cache or if the authentication credential does not match the cached credential. Kev --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] java.sql.SQLException: JZ006: Caught IOException
I am having the following exception: java.sql.SQLException: JZ006: Caught IOException: java.io.InterruptedIOException: Interrupted system call Note: The application is running on Solaris box. On Windows environment this doesn't happen. I am attaching the stack trace for detailed reference. Any ideas why this happens???!!! Could be that the system call is receiving an EINTR from the kernel, are you using a JDBC driver with native code? Have a look at the truss command and use it to see what jboss is doing, it may shed some light if this is the case. Kev --- This SF.net email is sponsored by: SF.net Giveback Program. SourceForge.net hosts over 70,000 Open Source Projects. See the people who have HELPED US provide better services: Click here: http://sourceforge.net/supporters.php ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] PB: The system cannot find the file specified !
ntrollerEJBBean\1061985263872.ser (The system cannot find the file What can be the reasons, that such a file cannot be retrieve? If the passivated bean has not been accessed for some time it is cleaned out by the system. Have a look at standardjboss.xml for the container configs of stateful session beans. We are running on a modified version of 3.0.2 and we came across this when the session bean could not be passivated because of a serialisation issue. The bean is just thrown away but the system believes it has passivated it and tries to activate it as normal. Kev --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Possible bug: unchecked does not seem to work.
Does not JBoss support the unchecked element in ejb-jar.xml/method-permission? You still need to have a principal though. Some login modules allow you to specify the principal that is used when unauthenticated access is requested. If you don't specify this then all access must be authenticated. Which modules are you using? Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by: The Definitive IT and Networking Event. Be There! NetWorld+Interop Las Vegas 2003 -- Register today! http://ads.sourceforge.net/cgi-bin/redirect.pl?keyn0001en ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Axis JBoss - Client Stubs Not Portable ?
Hiya Nicholas. I have been working with Axis JBoss for a while and I am finally getting around to putting together my build system. It has occured to me that when I run WSDL2Java, the client stubs generated have HARD CODED IP addresses or host names, making the generated code completely non-portable to other systems. So much for code once, run anywhere. The components affected are the ServiceLocator and the SoapBindingStubs. I think the only one which makes it unportable is using the default method in the locator to create a stub as this is used to communicate with the server. The url used is the endpoint url or, if not supplied, then the location generating the wsdl. The only ones I can think of in the stub are the type and operation namespaces, neither of which really have any functional bearing (they just partition the types and operations). The ServiceLocator issue seems to be trivial to fix, but the SoapBindingStubs use the IP address/host name in the namespaces. Use the methods with a URL parameter to provide the location of the service. If you don't like the generated one then you can override it using the endpointURL element within service (in the wsdd). The type and operations are affected by a number of things but the best one is probably the default namespace. This can also be specified in the wsdd file using the namespace elements in service. (The first namespace is the default one). Am I doing something wrong here ? Does anyone have any insight into this ? Let me know how you get on. Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by:Crypto Challenge is now open! Get cracking and register here for some mind boggling fun and the chance of winning an Apple iPod: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBoss.NET and Axis SOAP Monitor
How can I use Axis SOAP Monitor with JBoss? Axis site says it should be in http://localhost:port/axis/SOAPMonitor, but there is nothing. And does anyone know what is wrong with jboss.org site? It seems to be down.. You have to set it up yourself. You need to add a servlet mapping into your web.xml for the SoapMonitorService (I think), add the SoapMonitorApplet into the root of your web app and specify the handlers on the service (or transport). I'm not sure how this will work given the classloading in jboss but you can certainly try it. The soap monitor service creates a server socket to which the applet connects. A better way to see what is happening is to use the tcpmon as a proxy, it's easy to set up the HTTP communication to proxy through it. The tcpmon will show you all the headers (authentication info) as well as any error response (the soap monitor does not handle faults). It also means you can do it on a per client basis if you wish. HTH, Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by:Crypto Challenge is now open! Get cracking and register here for some mind boggling fun and the chance of winning an Apple iPod: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Remote Beans as Web Services
a) I looked through the JBoss.net EJBProvider code again. All we do there is to switch the classloading context and then delegate into the parent (Axis) EJBProvider. If there is a problem with * and wsdl generation, this should appear in plain Axis, too. I think I saw the right stop classes there. Kevin, have you any experience with that? We're using a modified version of Axis 1.0 within JBoss (because of a deployment decision) but from memory the EJBProvider didn't do this. ... I have just checked the 1.0 and 1.1rc2 source. The 1.0 source does not set these and the 1.1rc2 has it commented out. Which version are you using in jboss.net? Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by:Crypto Challenge is now open! Get cracking and register here for some mind boggling fun and the chance of winning an Apple iPod: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Remote Beans as Web Services
I'm not sure which version of Axis I'm using to be honest as it came installed as part of the embedded Catalina I downloaded. Is there any way of telling from the axis.jar which version I'm using? What attibutes does the 1.0 source not set? Sorry Mark but I think I've jumped onto the thread by mistake, I was replying to a comment made by Christoph. The problem I was talking about is connected with exposing all the methods in the soap interface, adding the ejb object interfaces to the stop classes list fixes this. This shouldn't cause the null pointer exception you are seeing. Sorry, Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by:Crypto Challenge is now open! Get cracking and register here for some mind boggling fun and the chance of winning an Apple iPod: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Remote Beans as Web Services
I'd a look at the EJBProvider code myself and the strange thing was that the line pointed to in the stack trace where the NullPointerException occurred was blank? I figured it maybe somehow it was the superclass method i.e. the axis EJBProvider that was maybe throwing the error somehow but sure enough there was nothing at that line either? The jars I'm using were from the embedded catalina version I downloaded from Dr. Scharlau's site. Hence why I'm trying to get Jboss running within Eclipse so I can see exactly what's causing this NullPointerException. It sounds as if jboss is picking up a different version of the library in that case, you would not expect to see an empty line unless the code differed from the compiled class. When you start jboss and connect to it remotely can you then step thru the code and insert the likes of break points etc? Yep, this can be done very easily. All you have to do is connect from eclipse as a remote debugged application. Can you send me the link to Dr. Scharlau's site? I seem to remember it being Aberdeen Uni. but can't find the URL. If you can also send me the backtrace and your jboss setup I'll have a quick look. Please send these direct and not to the list. Thanks, Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by:Crypto Challenge is now open! Get cracking and register here for some mind boggling fun and the chance of winning an Apple iPod: http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0031en ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] CMP column indexing
i was wondering if it's possible to specify indexing of a column in a CMP Entity Bean? I submitted a patch to do this for 3.0.2 but unfortunately I also included code to quote tables and fields. This was fine with the database we used but not with hypersonic. (actually it worked after I tidied up some of the meta data usage) I looked into the problem and discovered that hypersonic did not like quoting table names. I've been meaning to fix this and resubmit it, adding a flag to say whether quoting should be used for table names/fields etc. I'll pull down the latest 3.0.X code and see how easy it will be to update. We haven't used 3.2 nor 4.0, they may have this feature now. Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] CMP column indexing
when you say a patch - is the spec meant to allow this? I couldn't find it. Or is it a jboss specific extension? It is jboss specific, the spec doesn't mandate this as it is a performance optimisation specific to CMP mechanism. I had though about adding automatic index creation into the CMP engine but haven't had the time so far, I may get around to doing this at some point. Also, if anyone else knows about whether or not the later versions have this as a working feature I'd much appreciate it. It would save me some work :-) Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBoss.net and holders?
For one, if you have structured types that you send around (most likely bean-like), then you have to register particular de/serializers for them in the web-service.xml (see the jboss.net testsuite which should contain an xdoclet-enabled HelloData JavaBean in the hello chapter that is passed around by the Hello EJB). I didn't think that you had to register a de/serializer for java beans, doesn't the Axis code check for this and uses the bean one one if the class is a bean? I can't remember if you have to call setDoAutoTypes in the type mapping before using this (we have our own type mapping that uses the BeanDe/serializers.) Call-by-reference emulation through returning changed argument values can be implemented using custom serializers, but is IMHO not supported by Jboss.net or Axis or even WSDL per default. Axis does support the notion of IN/OUT and OUT parameters, which is where these holders are used. There is code in RPCProvider's processMessage which appears to handle the creation of these holders, for IN/OUT and OUT parameters. Having said all that I do not allow the specification of these parameters so I have never exercised that code. My emitter automatically generates the OperationDesc structures and forces the parameters to IN. The generated WSDL should contain a response message containing more than one part. I believe the name of the IN/OUT parts should match the parts in the request message. I'm not sure how this is handled in the generated java, presumably the client also has a Holder that has its valud modified. BTW Christoph, did you get a chance to look at the exception ideas I sent a few weeks ago? Thanks, Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JAAS authentication with server side accounting
Yes this was my first idea as well. I know I can implement a JAAS login module for the client side which will get invoked under and circumstances. But I need to account for the login failures on the server side. I don't want to implement a client side JAAS login module which will account for failed logins and than invoke a unsecured bean method so that these information get stored on the server side. This would just be to unsafe. I want to account for the login failure where they get checked on the server side. Secure your beans by assigning a JAAS security domain in jboss.xml. This will cause the server to invoke the associated JAAS modules and allow you to achieve the intended result. The client side JAAS module only needs to store the credentials and pass them over to the server, the supplied jboss one does this. The only thing to watch out for is that jboss will cache the credentials if the login has been successful. It will then compare these credentials with the supplied ones every time a bean is accessed instead of calling the JAAS modules, for reasons of performance. It will only call the JAAS module if the supplied credentials are different from the cached or if the cached values have expired. This is only an issue if the login credentials are changed (ant not propogated from the client) or you wish to block a running user. This caching behaviour can be turned off though. HTH, Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by: Scholarships for Techies! Can't afford IT training? All 2003 ictp students receive scholarships. Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more. www.ictp.com/training/sourceforge.asp ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JAAS authentication with server side accounting
Hiya Sebastian. So you suggest that I do something like this: 1. Authenticate the user with a client side JAAS login module We are not actually performing any client side authentication as such, we use the jboss JAAS login module solely to capture the user's credentials. The captured information is passed by the client interceptors to the jboss server when the beans (session or entity) are invoked. Our client does not know if the credentials are valid until it calls a bean method. 2. In any case invoke a arbitrary bean method on a bean under a security domain to get the server side JAAS login module invoked. Yep. At this point the jboss server invokes our real JAAS module and we can authenticate the credentials and prevent access if necessary. This works for us because our client calls a session bean to obtain user specific information before proceeding with any of our main screens. If this fails (it is an unchecked method) then the client knows that the credential are invalid and that the request was not denied because of the user's roles. 3. Account for the login failure using the server side JAAS module. Is this right? Is the bean invocation the only way I can get a server side JAAS module involved for authentication? It's the only way that I am aware of. Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.net email is sponsored by: Scholarships for Techies! Can't afford IT training? All 2003 ictp students receive scholarships. Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more. www.ictp.com/training/sourceforge.asp ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: AW: [JBoss-user] How to throw AxisFault in JBoss.net?
I'm away to Paris this weekend but I'll send you some more information
on Tuesday when I return.
After a wonderful weekend (and not so wonderful week) here's the second
part :-). Sorry for the delay.
As mentioned last week I have my own version of the EJBProvider. My
provider catches exceptions that occur when the call is made to the
session bean and wraps them in an AxisFault before the JavaProvider
steps in.
I have modified the wsdl2java generator to include code on the client
side to unwrap these exception and throw them if they are expected.
The server side looks like this: -
final Object serviceObj =
((OrchardEJBSessionServiceLifecycle)obj).getServiceObject() ;
try
{
return super.invokeMethod(messageContext, method, serviceObj, args)
;
}
catch (final InvocationTargetException ite)
{
throw makeAxisFault(ite, messageContext) ;
}
where makeAxisFault serialises the exception and creates an AxisFault with
the result as a detail element.
The client side looks like this: -
try
{
java.lang.Object _resp = _call.invoke(new java.lang.Object[] {in0});
if (_resp instanceof java.rmi.RemoteException) {
throw (java.rmi.RemoteException)_resp;
}
}
catch (final org.apache.axis.AxisFault fault)
{
final Throwable origException =
uk.co.orchardsystems.soap.client.OrchardSerialisationHelper.deserialiseFault
(_call.getMessageContext(), fault) ;
if (origException != null)
{
if (origException instanceof
uk.co.orchardsystems.security.client.OrchardSecurityCreateException)
{
throw
(uk.co.orchardsystems.security.client.OrchardSecurityCreateException)origExc
eption ;
}
if (origException instanceof
uk.co.orchardsystems.security.client.OrchardSecuritySystemException)
{
throw
(uk.co.orchardsystems.security.client.OrchardSecuritySystemException)origExc
eption ;
}
if (origException instanceof java.lang.SecurityException)
{
throw (java.lang.SecurityException)origException ;
}
}
throw fault ;
}
The soap looks like this: -
?xml version=1.0 encoding=UTF-8?
soapenv:Envelope xmlns:soapenv=http://schemas.xmlsoap.org/soap/envelope/;
xmlns:xsd=http://www.w3.org/2001/XMLSchema;
xmlns:xsi=http://www.w3.org/2001/XMLSchema-instance;
soapenv:Body
soapenv:Fault
faultcode
xmlns:ns1=http://xml.apache.org/axis/;ns1:Server.userException/faultcode
faultstringAuthentication exception, principal=orchard/faultstring
detail
ns1:exception xsi:type=ns1:java.lang.SecurityException
xmlns:xsi=http://www.w3.org/2001/XMLSchema-instance;
xmlns:ns1=http://www.orchard-systems.co.uk/wsdl/types;
message xsi:type=xsd:string
xmlns:xsd=http://www.w3.org/2001/XMLSchema;Authentication exception,
principal=orchard/message
/ns1:exception
/detail
/soapenv:Fault
/soapenv:Body
/soapenv:Envelope
The only part I am unsure about is the format of the soap request
but I believe it is correct from my reading of the spec.
The other thing to watch out for is that the axis code has a bug that
does not correctly generate the wsdl for exception classes nor serialise
them properly. I have had to replace their serialiser/deserialiser with
one of my own. I can shed more on this later if it has not already been
fixed.
HTH,
Kev
Kevin Conner
This is a personal e-mail. Any views or opinions presented are
solely those of the author and do not necessarily represent those
of Orchard Information Systems Ltd.
---
This SF.NET email is sponsored by: FREE SSL Guide from Thawte
are you planning your Web Server Security? Click here to get a FREE
Thawte SSL guide and find the answers to all your SSL security issues.
http://ads.sourceforge.net/cgi-bin/redirect.pl?thaw0026en
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: AW: [JBoss-user] How to throw AxisFault in JBoss.net?
So would it be possible to write a special request handler in the bean´s transport chain that takes your (non-Axis-Fault, means that the container will let it through unchanged) application exception and wraps it into an appropriately massaged axis fault such that the name of the original exception will appear in the response envelope? Unfortunately not. The JavaProvider, if my memory is correct, catches the exception and changes it into an AxisFault before it gets back to the chains. But even if it did get back to the provider chain it would still not get back to the transport one, at least not with Axis 1.0. The onFault method is only called on the handlers in the current chain, the AxisServer allows the AxisFault to continue unimpeded once the current chain has handled it. I'm not sure if this is the intended behaviour but it is certainly the implemented one. I would have hoped for something more like the network streams or jboss' invocators. Then you can have a client-side stub that reconstructs the exception for you I have my own EJBProvider that dynamically exposes a group of our session beans as ports within a service. When this calls the invoked method it catches the InvocationTargetException and wraps a serialised version of the exception in an AxisFault. The generated proxy code then checks the axis fault to see if it contains one of the expected exceptions and then throws it. I also had to tweak the generators to prevent generation of the derived axis faults. I'm away to Paris this weekend but I'll send you some more information on Tuesday when I return. Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: AW: [JBoss-user] How to throw AxisFault in JBoss.net?
Sorry to take so long in responding but I've just been going through this issue in our code, stand alone axis 1.0 + jboss (not jbossnet). This approach is outdated, because the best way, IMHO, is to take the server-generated wsdl and pre-generate client stubs using the wsdl2java toolkit. They should have some exception support there ready to remap the AxisFault to ordinary java exceptions, shouldn´t they? Classname is there, message is there, maybe not the stack trace. The 1.0 code only adds the exception name if it is a subclass of AxisFault so only the axis faults can be reproduced on the client side. All other exceptions get translated into an AxisFault containing only the message and stack trace. I have had to extend the wsdl2java to include my own type mapping hierarchy and exception handling (the type mapping for some custom DTOs that are not java beans). The exception gets serialised into XML on the server side and placed into the soap fault detail by my own EJB provider. The new generated proxy code catches the AxisFault on the client and, if the exception detail is present, regenerates the original exception and throws it instead. This has required a very big hack to get it working but it does work. The changes are very fragile so I would be loathed to change to another version of Axis without a lot of work. It is certainly safer to stick with the Axis implementation but unfortunately that was not an option open to me. Kev Kevin Conner This is a personal e-mail. Any views or opinions presented are solely those of the author and do not necessarily represent those of Orchard Information Systems Ltd. --- This SF.NET email is sponsored by: SourceForge Enterprise Edition + IBM + LinuxWorld = Something 2 See! http://www.vasoftware.com ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] How to log in from an MBean
What I am saying is, since I have declared a domain called myDomain using the user database of my system, I think it is reasonable to expect that both the Swing client and the MBean client could log in using the myDomain. As mentioned, if my MBean carries out a login on the myDomain using, say, the same credentials as my Swing client, the log in is of course successful, however, due to the way JBoss works my MBean still does not get access to my EJBs. For my MBean to get access to EJBs it has to use a different log in module configuration, namely the client-login referring to the ClientLoginModule. But this is required for all client access, whether it comes from a swing client or not. The ClientLoginModule does not do any authentication, it is used by JBoss to assign security credentials to the current thread (or application in the case of your swing client). This information is then retrieved from the context by the invokers when they call the EJB. You can still use your own authentication module but it must be WITH the ClientLoginModule. If you do not use ClientLoginModule then JBoss has no security context to pass via the invokers. Unfortunately I have not yet contributed code to JBoss, so I cannot tell whether there is a sound design reason for the way this is. All I am saying is, I find it reasonable for JBoss users to expect that when a Swing client can log in and get access to the EJBs using myDomain, so can any other client - including MBeans. And that should happen if they are both configured the same. Are you certain the credentials obtained on your client side are the ones you are seeing on the server? Is it possible that you have called the ClientLoginModule via another domain? If you are not in a server environment then the same security context is assumed for every thread (unless you specify the 'multi-threaded' option). The technical reason is that JAAS does not provide any way of associating the security context with the current thread (or application). The ClientLoginModule allows JBoss to capture this information and store it for later use. HTH, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBoss component library revisited
o Re-use of existing component libraries etc. As other people on the list have protested, why re-invent the wheel? Of course, life is rarely that simple. Using other component libraries etc. is likely to only be possible if they follow the same standards, licenses, philosophy etc. My experience has led me to believe that software tends to be less re-usable than we pretend it is. Yep, and if everybody's response is 'lets write our own' then that will always be the case! Fix the software already in place! What is the point of having multiple component libraries? If there is something you dislike in the current libraries then work with the developers to fix the issue. There is no point in dividing valuable resources just because the current offerings aren't quite what you want. Work with the others not against them. Where would JBoss be now if everyone decided to develop a new app server just because there was something 'not quite right' with it? JBoss has evolved into its current offering because people have taken the time to use it and offer feedback. They have worked with the core developers to improve the product and not tried to compete with it. Please do the same with the libraries. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBOSS - Components library
Sorry to start this again, I've been away. 1. to showcase Jboss (yes, they could run on other servers, but focus on JB as the 1st server) 2. provide a set of components that could be used as-is or customized for your application 3. provide some example Uis either using a rich or thin client that utilize these server-side component 4. provide a good jumpstart for new applications with a starter build system (combine ant, xdoclet, junit, etc. etc.) , several prewritten applications that use the apis as examples or skeletons, etc. If the purpose of this is to showcase JBoss then this would get my vote. If it is to create a new library then it wouldn't (for reasons mentioned earlier). When I contemplated embarking on a simliar OSS project that would work on the top 5 app servers, I got some interesting comments from Rickard Öberg, which I neglected to pursue with further discussion at the time (shame on me): Let's say I wanted to impose access checks on modifying a user, or if I wanted to add custom fields to an account, how would I do that with your solution? Do you have any way to handle extensibility and integration? How about getting user info from various backends, such as RDBMS or LDAP or XML or..? IMHO there are vital points missing in todays toolkit in order to allow wide reuse of components such as those you outline. These holes will be filled, but IMHO they're not right now. I think it would be a very good idea to show off the capabilities of JBoss, after all the easiest way to convince someone to use it is via demonstration. We should, however, use some of the projects which are already out there as a basis. If they do not have what we require then we should persuade them to change/add functionality from the inside (as members) or outside (as users). We would then be in a position to say 'yes you can do this with another server but look what else you can do with JBoss!' So, I would think the idea would be to ensure its OSS so we can make changes to the code, encourage it for new projects where you may be able to have some flexibility in your implementation if something is currently missing from a component's API, and not get too fancy with things like XML-code generators and other things since there are projects out there that provide these solutions already (see previous messages in this thread). KISS. James A second for KISS. Please do not reinvent the wheel! Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: DEDICATED SERVERS only $89! Linux or FreeBSD, FREE setup, FAST network. Get your own server today at http://www.ServePath.com/indexfm.htm ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBOSS - Components library
That is my point exactly - showcase Jboss, build as much using the standard j2ee approach, integrate interesting libs as needed, and make something professional. This may mean trying to find someone who is a graphic artist to put together a nice look for the examples. Anyone can write an ugly demo.. This demo and the associated components should be something you would go to a tech conference with, if the purpose is to woo potential converts.. Just ask yourself, What would BEA do? Now, maybe I'm going overboard and this is something the Jboss services team has done or should do. But one thing I have learned - if you are targeting business (and even some tech) folks to convince them to go Jboss, sharp Uis win out most of the time.. Jboss has it where it counts - architecture, now let's take it to the next level - win the business folks as well. So, two tracks: demo the tech capabilities of Jboss with demos, and make it sharp for the business folks that need convincing. Thoughts? Sorry, I included your comments because I agree with them. I also wanted to stress my original objection to creating a NEW components library. I agree completely with what you are saying. We should be showcasing JBoss by building upon an existing component framework and, as much as is possible, improving that framework instead of creating yet another. It is true that some things may be JBoss specific but we should aim to keep that to a minimum. Sorry for the confusion, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: DEDICATED SERVERS only $89! Linux or FreeBSD, FREE setup, FAST network. Get your own server today at http://www.ServePath.com/indexfm.htm ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBOSS - Components library
As an alternative, we could use an existing (EJB) component sharing forum - any suggestions? There are a number of them out there, why start another? Two that spring to mind are ofbiz (Open For Business) and Open Symphony. I've never used either of them so I don't know what their quality is like nor how complete they are. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBOSS - Components library
This just what I talked about, to centralize the development of such components, I mean *FREE* components working without spending one hour to configure it... sort of Plug and Play components ;) I was talking about *FREE* components. For Open For Business, it's not the same, here I'm talking about Business Components. I thought ofbiz came with business components. They certainly provide core services such as workflow, security, rules engine etc. but I thought it also came with domain specific code. It's obviously time for me to have a look at it again. For OpenSymphony, it's near my idea but it don't seems to be very developed. OpenSymphony wishes to provide components for any App. Server. Here, I'm talking of JBoss only (at least for now). There are a lot of companies out there who cannot choose to use jboss, why tie these components to jboss? Why do a M$ on this? Perhaps, a JBoss Component Library could start with such projects... but I still think no appropriate solution exists for now. Or perhaps such solutions should have a place on the website jboss.org... (the truth is out there... :) ) Then go to those other projects and suggest the changes you would like to see. JBoss has become so good because they are forward thinking and very receptive to ideas (and help) from outside, why do you think the others will not be the same? Why reinvent the wheel? If you have better ideas then give them the chance to fit in with them. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBOSS - Components library
Well, I've had a quick look and saw the Entity engine, not entities ready to use. From memory- The entity engine is for managing data definitions, not entities as in EJBs. I think they have a generic entity (DTO) and define the contents/tables using XML. Well, I think: no one than the JBoss's developers can make better components. But I know, that's not true. I say that because I'm bored to spent time and time to develop just an EJB. What I want to show is that I'm not he only one who stands in this situation :( If such a sub project starts, and is managed by the JBoss group/team/..., I intends to participate (at least to do what I can in this way). I feel very concerned by the open source :) There are some other projects out there who have very good developers and, as good as the JBoss guys are, some of the others will be better :-). They could also be worse but that's where we come in :-) Totally agree! I didn't want to tell other projects are bad. I just want to tell: Work together, and prove to the world, it must count with Open Source :) I could add Work together, with standards... ;) Here here! Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: AW: [JBoss-user] jboss.net and persistence
after doing some research (trial and error mostly) it seems that I also need to add: parameter name=scope value=Session/ to the service element of my web-service.xml file to get jboss.net to remember I want to maintain a session. Have to recheck that, I´m not sure where the ejb reference is put by axis, but it is a good hint. The default scope in Axis is 'Request'. If you do not change this then the service will be created and destroyed every time a request is serviced. 'Session' and 'Application' are the others. I don't know if JBoss.net has a different EJB provider, we're bundling Axis within our ear file, but the Axis provider requires you to specify a number of options that can be determined from the EJBMetaData :-( Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: AW: [JBoss-user] jboss.net and persistence
We have subclassed the Axis EJBProvider to only depend on the JNDIName (and the Scope as it seems). That's what I've done but I must admit that I kept it simple and only used the remote interface :-), I haven't added anything for local interfaces. I would need to examine the create methods and use the return type to get the local interface. Actually the more I think about it the more I like it. I think it's time to revisit my provider and add this. BTW I also changed the return type to use the ServiceLifecycle interface for notifications. This allows me to call remove when it goes out of scope. I also had to add a listener to the servlet session as Axis doesn't cope with this yet. Unfortunately, EJBMetaData is only accessible from EJBHome and not from EJBLocalHome (which makes sense when designing an application that just uses in-VM RMI communication and out-VM SOAP). I hesitate building two separate Providers, especially since the automatic scoping behaviour would be only available for one of them. Hmm. But what about having the xdoclet template caring about that scope issue? I guess the real question then becomes why wouldn't you have a session bean with a minimum scope of 'Session'? The only reason I can think of is to keep the number of cached objects to a minimum. Kev P.S. The only reason we are bundling Axis within our ear is that we intend to deploy on WebSphere when they eventually release it, unless I can convince my boss to use JBoss :-). Of course if that does happen then I can stop adding to Axis because you'll already have done it :-). Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Serious problem with wrong data returned.
If you have a simple test then we can try to reproduce it. /** * ejb:interface-method */ Which version of xdoclet are you using? We found a bug in version 1.1.2 when generating primary keys, it manifests itself during serialisation. Try regenerating your pk with the attached template. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 entitypk.j Description: Binary data
RE: [JBoss-user] Serious problem with wrong data returned.
Actually, it did not help. But thank you very much anyway, because it was in the entitypk.j the error occured. The problem was that our primary key was in a superclass, but the entitypk.j was not able to generate correct equals method for the subclasses. (All primary keys was equal...) Glad I could help, even if it wasn't what I thought :-). We haven't come across that one because our primary keys are in the subclass, this may change soon though. Thanks for the warning. BTW How did you fix it? Did you change the hashCode, equals and toString XDtEjbPersistent:forAllPersistentFields superclasses attribute to true? Thanks, Kev P.S. Watch out for the hashCode one because it does cause problems when serialising. Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] JBoss 3.0.1 and Oracle-XA
I am trying to get Oracle-XA to work with JBoss 3.0.1, by doing the following: 1. I copy the docs/examples/jca/oracle-xa-service.xml 2. I change the line with URL=jdbc:oracle:oci8:@tc to URL=jdbc:oracle:oci8:@banklab3:1521:ora8 3. I insert username and password 4. I include an attribute name=Padtrue/attribute into jboss-service.xml It looks like the mbean is registrated, but I get error message that complain with no ocijdbc8 in java.library.path I assume no 2 above is not correct. But how should the URL be then? The error message is saying that a required shared library, ocijdbc8.dll (or libocijdbc8.so), is not in the library path. You'll need to find it (and probably lots of others) and place them on the path (or LD_LIBRARY_PATH). One place to put it is in the jboss bin directory but this isn't nice. I believe you can also wrap all this in a sar file but I haven't tried yet. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] RE: Indexing of finders
Hiya Dain. That is me. I like this feature. Are you thinking about just adding an empty tag to the cmp-field element. Something like index/? I've managed to do the first part, specifying the indexes in the jbosscmp-jdbc.xml file. I've included the changes made to the 3.0.2 download for you and the others to have a look at (sorry, that's what we have at present). Let me know what you think and tell me if there are any glaring problems with my code or assumptions. The bit I'm worried about the most is the executeSql method in JDBCStartCommand. I've written this to take a list of SQL statements and execute them within a single call, using batching if present. If I've understood the rest of the code in that file then the connection is being used outside of a transaction context. I have therefore turned off auto commit and explicitly used commit/rollback. Is this assumption correct? BTW I've only tested this on the hypersonic database which comes with jboss, I would like to hear from others about any problems. I have tried to keep the coding style of each part the same as the style used in each individual file (hence the different styles). I have occassionally lapsed into my standard style, executeSql is an example. I've attached a small example of the jbosscmp-jdbc.xml file to give people an idea of how they are specified. I intend to include the auto generated indexes at a later date, hence their inclusion in the code. I may not get a chance to start this for a few weeks though as I'm in the middle of packing to move house. BTW Which CVS branch should I be using as a base for this? I'm going to make the changes to our local source initially, which has drifted from head, and then will look at moving it into CVS. It's just so that I can try out the changes before updating our version of jboss. Now that 3.2 is out I'll have a look at basing this code on that release. Kev P.S. I've included both mailing lists so that we can move this from jboss-user to jboss-devel. Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 jbosscmp-jdbc.xml Description: Binary data diffs Description: Binary data
RE: [JBoss-user] RE: Indexing of finders
I think he's talking about getting JBossCMP to generate a 'CREATE INDEX' statement at the time of table creation, to optimize the database end of the finder. speculation, but that's something I'd want. Except that it never occured to me and I always just added the indexes manually. Oh, well. That's exactly what I want. CMP is handling the creation of the tables and appropriate accesses but it does not seem to be indexing the fields used in finders. I know this is not part of EJBQL but then I never expected it to be, after all its job is to specify the finder logic not the implementation details. This is CMP engine specific, something which helps the particular CMP engine to execute the finder quickly. Do I understand from the rest of the thread that this is not currently possible with JBossCMP? If so, is there any interest in having it added? I'm willing to give it a go if the person responsible for JBossCMP is available to bounce ideas off. Is that you Dain? Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- In remembrance www.osdn.com/911/ ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] Indexing of finders
I have a simple question which I hope will have a simple answer :-) I have been going through the jboss 3.0.2 source to try and find out if there is any way to enable indexing for finder methods. So far I have been unable to find any mechanism for achieving this. Can anyone tell me if this is possible and, if so, what I have missed in enabling it? Thanks, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] RE: Indexing of finders
I have a simple question which I hope will have a simple answer :-) I have been going through the jboss 3.0.2 source to try and find out if there is any way to enable indexing for finder methods. So far I have been unable to find any mechanism for achieving this. Can anyone tell me if this is possible and, if so, what I have missed in enabling it? Sorry, forgot to add that this is for CMP entity beans with finders written in EJBQL. Thanks, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] eclipse with xdoclet and jboss
The stacktrace shows: javadoc: In doclet class xdoclet.DocletTask$DocletMain, method start has thrown an exception java.lang.reflect.InvocationTargetException java.lang.NoClassDefFoundError: org/apache/tools/ant/types/EnumeratedAttribute When I check the ant.jar, EnumeratedAttribute is there, and ant of course starts fine, so what's the problem? Yep, it's javadoc complaining. The initial classpath in eclipse does not contain the ant libraries and it is this classpath that is passed to javadoc. I bodged this by changing the antsupport library in eclipse to pass through the classpath. I remember changing AntRunner and InternalAntRunner but I will need to check what changed. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] eclipse with xdoclet and jboss
Yep, it's javadoc complaining. The initial classpath in
eclipse does not contain
the ant libraries and it is this classpath that is passed to
javadoc. I bodged
this by changing the antsupport library in eclipse to pass
through the classpath.
I remember changing AntRunner and InternalAntRunner but I
will need to check what
changed.
One thing I forgot to add. You do not need to bodge the eclipse
ant library to get it working. The other way is to include the
ant jars in the xdoclet classpath, i.e. something like: -
path id=xdoclet.classpath
fileset dir=${xdoclet.lib.dir} includes=**/*.jar/
fileset dir=${ant.lib.dir} includes=**/*.jar/
pathelement path=${java.class.path}/
/path
taskdef name=ejbdoclet classname=xdoclet.ejb.EjbDocletTask
classpathref=xdoclet.classpath/
...
ejbdoclet classpathref=xdoclet.classpath
/ejbdoclet
Please try a variation of this, it has been copied by hand from
one of my build files. :-)
Good luck,
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA. United Kingdom
Registered in England, Number 1900078
Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515
---
This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone? Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390
___
JBoss-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] behaviour of a cached initialContext
And about reusing homes (will JAAS principal be maintained?) The JAAS principal is not associated with the homes, you can safely cache them. The principal is associated with the current thread when the request is received. This thread local principal is then used by the proxy classes when the home (or bean) is accessed. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: Jabber - The world's fastest growing real-time communications platform! Don't just IM. Build it in! http://www.jabber.com/osdn/xim ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] jboss custom config not working from standalone client
but it works fine when i set the bindaddress ip to localhost. I think the problem is at jboss bindaddress conf. implementation. Any thoughts? java.rmi.UnmarshalException: error unmarshalling arguments; nested exception is: java.io.StreamCorruptedException Can you provide any more information on the backtrace? Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] jboss custom config not working from standalone client
put a bindaddress element like the following in jboss-servicexml, in the org.jboss.web.WebService and org.jboss.naming.NamingService mbeans : attribute name=Port8083/attribute (ex from Webservice mbean) attribute name=BindAddress10.1.11.116/attribute use your address for this instead of 1.1.11.116. Try this with 3.0.0 and with 3.0.1. With 3.0.0 it will work, and with the later not : ) I've already done this with 3.0.1 and it worked perfectly, I actually tried it before sending the previous reply. :-) I even verified it in the debugger, using netstat and with a zonealarm prompt. Can you give me more information about the backtrace? If I can get an idea of what you are doing and where it is going wrong then I may be able to help. Thanks, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] jboss custom config not working from standalone client
thanks for the help kevin. Your welcome, although I haven't done much yet :-) I'll install from scratch and test again, maybe i made some fault at somewhere... If it still fails please capture the backtrace. This may shed some light on the problem and may indicate where my test differs from yours. Good luck, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] jboss custom config not working from standalone client
I setup a customized configuration of final jboss 3.0.1 with the following files: jbossmq-destinations-service.xml jmx-console.war jbossmq-service.xml oracle-service.xml ejb-management.jar jca-service.xml tomcat4-service.jar jboss-local-jdbc.rar jms-ra.rar tomcat4-service.xml jboss-xa.rar jms-service.xml user-service.xml I use this confuration to allow more than one instance of jboss at same time (in a central machine for all developers). The problem is that the tests from junit (standalone client) doesn't work. it gives the message : java.rmi.UnmarshalException: error unmarshalling arguments; nested exception is: java.io.StreamCorruptedException I use a bindaddress in the jboss-service of custom config to avoid port conflict. The custom config worked in 3.0.0. We saw this problem when we moved from 3.0.1RC1 to 3.0.1 and traced it down to a 3.0.1RC1 server being called from a 3.0.1 client. Further investigation revealed that the externalised representation of org.jboss.invocation.MarshalledInvocation had changed between the two. Our problem was fixed by upgrading the server to 3.0.1. If you are seeing the same then it is probably because your client and server libraries are mismatched. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] MBeans not deploying
There has been some talk over the last day about mbeans not deploying for various reasons, I have another one to add :-) Our setup has jboss 3.0.1 jetty acting as a netboot server to developers machines, again 3.0.1. I have configured the client jboss servers to use the URLDirectoryScanner and specified common services using specific URLs. All configurations are correct (note I am not talking about the security name issue). When the scanner is checking for deployment it uses the URLConnection getLastModified method, which sends a GET request to the server. Sometimes this GET request is closed without any information coming back (I haven't chased this part up yet) but the result is that the lastModified time reported to the scanner will be 0. If this happens on the first scan then the lastModified date will match the default date of 0 and therefore the mbean will not be deployed. If this happens on a subsequent scan then the mbean will be undeployed and then redeployed. This can happen to any mbean. I spent yesterday chasing this up in the code and have verified it using a network snooper. I am now turning my attention to the jboss netboot server to see why it is not handling the requests. If anyone has an idea then please let me know, it will save some time on my part and allow me to continue the development work I am supposed to be doing :-). Kev P.S. I haven't seen this happen on the server archives but presumably it can. Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
[JBoss-user] RE: MBeans not deploying
I spent yesterday chasing this up in the code and have verified it using a network snooper. I am now turning my attention to the jboss netboot server to see why it is not handling the requests. If anyone has an idea then please let me know, it will save some time on my part and allow me to continue the development work I am supposed to be doing :-). Forgot to add the following. Server is MP Solaris 5.8 running jdk1.3.0_04. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] MBeans not deploying
Would it be possible to solve this problem by recognizing the communication failure, marking it with a special value in lastModified, and doing nothing until a real response was obtained? The only problem is that this isn't a communication failure. The request is getting to the server and the server is closing the socket without generating a response. The current getLastModified code does not differentiate between a FileNotFoundException (404 response), a general IOException and a successful connect without any response. The exists method is similar in that it may cause a mbean to be undeployed because of a communication glitch and not a 404 response. It might be better to change the code so that the getLastModified and exists can be rolled into one method, returning a does not exist (1), comms error (2), lastModified unknown (3) or lastModified time (4). It can then choose to ignore the information for 2 and 3 but not 1 and 4. This will then cope with some of the network's (or server's) foibles :-). Of course you may find that when you go to redeploy you will get hit again and will end up with an undeployed mbean. But you can't cope with everything :-). At least this window only happens when you believe there is a change to apply. Kev P.S. I'm still no further on the netboot server side, the problem disappears once I add a certain amount of debugging :-). Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] MBeans not deploying
It might be better to change the code so that the getLastModified and exists can be rolled into one method, returning a does not exist (1), comms error (2), lastModified unknown (3) or lastModified time (4). It can then choose to ignore the information for 2 and 3 but not 1 and 4. This will then cope with some of the network's (or server's) foibles :-). I've made a stab at this, changing the internal Scanner class for URLDirectoryScanner. The problem still occurs but it does not force a redeployment of the mbean. This is fine for our current purpose and I've rolled it out to our netboot environments. Of course this does not deal with the issue of what is happening on the netboot server side, it is still closing the connection without sending a response. I'll try to find some time to look at it at a later date :-). Can someone please take a look at the changes I've made? If they are acceptable can they be rolled into the current versions? If not can they be used as a basis for a fix? Thanks, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 URLDirectoryScanner.java Description: Binary data
RE: [JBoss-user] MBeans not deploying
Can someone please take a look at the changes I've made? If they are acceptable can they be rolled into the current versions? If not can they be used as a basis for a fix? Please ignore this request. I've decided that this is an inappropriate forum for and am logging a bug report instead. Thanks, Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: OSDN - Tired of that same old cell phone? Get a new here for FREE! https://www.inphonic.com/r.asp?r=sourceforge1refcode1=vs3390 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
RE: [JBoss-user] Netboot
If they lie on a central server, then he only sees the interfaces, but has no access to class files of the EJB implementations (please correct me if I'm wrong). He has the swing-app, but he has not the required EJBs. You could always place a delegate layer on the client and have it switchable between soap and session facades (or just use soap). The jboss server can service the soap calls and pass them through to the ejb layer. In this way the client will not even be exposed to your implementation technology. Kev Kevin Conner Orchard Information Systems Limited Newcastle Technopole, Kings Manor Newcastle Upon Tyne, NE1 6PA. United Kingdom Registered in England, Number 1900078 Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515 --- This sf.net email is sponsored by: Dice - The leading online job board for high-tech professionals. Search and apply for tech jobs today! http://seeker.dice.com/seeker.epl?rel_code=31 ___ JBoss-user mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-user
