Re: [JBoss-dev] HTTPSession Clustering: howto
Hey, On Mon, 31 Dec 2001, Sacha Labourey wrote: Servlet container implications == As we exchange an HttpSession object between the servlet container and the EJB, it needs to be Serializable. Consequently, a new interface has been defined: public interface SerializableHttpSession extends javax.servlet.http.HttpSession, java.io.Serializable { public boolean areAttributesModified (SerializableHttpSession previousVersion); } This interface makes sure we have a HttpSession compatible interface that is Serializable. Furthermore, a method has been introduce that ask the HttpSession to decide when 2 sessions are differents. As indicated in the Javadoc, this method should *not* check all data (such as the last access time). This in order to reduce the cluster communication (more on this in the JavaDoc comments of the JMX Service) Are you going to use this interface to decide if you need to re-replicate the httpsession on other nodes? If this is the case consider having a method that returns a list of modified attributes keys, thus enabling you to replicate only what you actually need to replicate rather then replicating whole httpsession again. You can also I guess use the method to check if httpsession is changed i.e: ((changed == true) == attributeChanged.size()0) Cheers, Vladimir ___ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] A Weird Question about Clustering
Hey Joseph, You raise very interesting point, I've breezed through the papers you reference and most of them are in either early stages of conception or an infancy. IBM cJVM http://www.haifa.il.ibm.com/projects/systems/cjvm/cjvm_papers.html This seem to be the most mature project but it for example doesn't mention the details of fast interconnection between the nodes. I assume it is n-(n-1) tcp model which is a nightmare in any case, Anyway, you organize your global heap, i.e put this many master objects on this node that many on another one, proxies on all other , in order to share the computational workload, but you end up remote calling across the wire in object communication. Maybe they have some algorithms that give the best trade off in distribution amount and communication minimization? The clustering model IBM offers is no doubt a perspective project but I think it fits more in science computation purposes rather than app server model. TIGRIS http://www.cs.berkeley.edu/~mdw/papers/tigris-lcr2k.pdf Matt Welsh always has some interesting things to say. I'll read this one in details. The group communication approach , the one that jboss currently uses is IMHO more suitable for app server since your are primarely concerned with object replication rather than thread migration and such things. If you are interested in group communication, there is tons of interesting research out there. You may start with prof. Birman's work at Cornell, prof. Babaoglu at University of Bologna, John Hopkins University and prof. Yair Amir, prof.Dolev from Hebrew University of Jerusalem , prof Moser at University of California at Santa Barbara, Bela Ban's Javagroups etc etc. Cheers, Vladimir ___ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmx/src/main/org/jboss/ha/framework/interfaces HAPartition.java HARMIClient.java
User: vlada Date: 01/10/01 21:05:56 Modified:src/main/org/jboss/ha/framework/interfaces HAPartition.java HARMIClient.java Log: rather than using machine dependent system time to track if client's view of cluster has changed use logical viewID from cluster Revision ChangesPath 1.2 +2 -1 jbossmx/src/main/org/jboss/ha/framework/interfaces/HAPartition.java Index: HAPartition.java === RCS file: /cvsroot/jboss/jbossmx/src/main/org/jboss/ha/framework/interfaces/HAPartition.java,v retrieving revision 1.1 retrieving revision 1.2 diff -u -r1.1 -r1.2 --- HAPartition.java 2001/09/30 15:19:32 1.1 +++ HAPartition.java 2001/10/02 04:05:56 1.2 @@ -20,7 +20,8 @@ public DistributedReplicantManager getDistributedReplicantManager(); public DistributedState getDistributedStateService (); - + public long getCurrentViewId(); + // *** // *** // RPC multicast communication 1.2 +7 -15 jbossmx/src/main/org/jboss/ha/framework/interfaces/HARMIClient.java Index: HARMIClient.java === RCS file: /cvsroot/jboss/jbossmx/src/main/org/jboss/ha/framework/interfaces/HARMIClient.java,v retrieving revision 1.1 retrieving revision 1.2 diff -u -r1.1 -r1.2 --- HARMIClient.java 2001/09/30 15:19:32 1.1 +++ HARMIClient.java 2001/10/02 04:05:56 1.2 @@ -30,7 +30,7 @@ * * @author a href=mailto:[EMAIL PROTECTED];Sacha Labourey/a * @author a href=mailto:[EMAIL PROTECTED];Bill Burke/a - * @version $Revision: 1.1 $ + * @version $Revision: 1.2 $ * * pbRevisions:/b * @@ -45,7 +45,7 @@ protected String key = null; protected ArrayList targets = null; protected LoadBalancePolicy loadBalancePolicy; - protected transient long tag = 0; + protected transient long currentViewId = 0; protected transient Object local = null; public HARMIClient () @@ -89,16 +89,7 @@ targets.addAll (newTargets); } } - - public long getTag () - { - return tag; - } - - public void setTag (long tag) - { - this.tag = tag; - } + public Object getRemoteTarget () { @@ -139,7 +130,8 @@ } // nothing found } - + + public Method findLocalMethod (Method method, Object[] args) throws Exception { return method; @@ -185,12 +177,12 @@ RemoteMethodInvocation rmi = new RemoteMethodInvocation (null, method, args); MarshalledObject mo = new MarshalledObject (rmi); // Is this step actually necessary? Can I just do method.invoke(target, args); ? -HARMIResponse rsp = target.invoke (tag, mo); +HARMIResponse rsp = target.invoke (currentViewId, mo); if (rsp.newReplicants != null) { System.out.println (new set of replicants); setTargets (rsp.newReplicants); - setTag (rsp.tag); + currentViewId = rsp.currentViewId; } return rsp.response; } ___ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmx/src/main/org/jboss/ha/framework/server HAPartitionImpl.java HARMIServerImpl.java
User: vlada Date: 01/10/01 21:06:47 Modified:src/main/org/jboss/ha/framework/server HAPartitionImpl.java HARMIServerImpl.java Log: rather than using machine dependent system time to track if client's view of cluster has changed use logical viewID from cluster Revision ChangesPath 1.2 +10 -3 jbossmx/src/main/org/jboss/ha/framework/server/HAPartitionImpl.java Index: HAPartitionImpl.java === RCS file: /cvsroot/jboss/jbossmx/src/main/org/jboss/ha/framework/server/HAPartitionImpl.java,v retrieving revision 1.1 retrieving revision 1.2 diff -u -r1.1 -r1.2 --- HAPartitionImpl.java 2001/09/30 15:20:58 1.1 +++ HAPartitionImpl.java 2001/10/02 04:06:47 1.2 @@ -198,7 +198,12 @@ { return this.dsManager; } - + + public long getCurrentViewId() + { + return channel.GetView().GetVid().GetId(); + } + // *** // *** // RPC multicast communication @@ -424,8 +429,10 @@ newMembers.add (allMembers.elementAt (i)); return newMembers; } - - + + + + /// /** 1.2 +5 -10 jbossmx/src/main/org/jboss/ha/framework/server/HARMIServerImpl.java Index: HARMIServerImpl.java === RCS file: /cvsroot/jboss/jbossmx/src/main/org/jboss/ha/framework/server/HARMIServerImpl.java,v retrieving revision 1.1 retrieving revision 1.2 diff -u -r1.1 -r1.2 --- HARMIServerImpl.java 2001/09/30 15:20:58 1.1 +++ HARMIServerImpl.java 2001/10/02 04:06:47 1.2 @@ -40,7 +40,6 @@ DistributedReplicantManager.ReplicantListener { protected String replicantName; - protected long lastSet = System.currentTimeMillis (); protected ArrayList replicants = new ArrayList (); protected Object handler; protected HashMap invokerMap = new HashMap (); @@ -100,12 +99,8 @@ e.printStackTrace (); } } + - public long getTag () - { - return lastSet; - } - public Object getLocal () throws Exception { return handler; @@ -121,10 +116,9 @@ replicants.clear (); replicants.addAll (newReplicants); } - lastSet = System.currentTimeMillis (); } - public HARMIResponse invoke (long tag, MarshalledObject mimo) throws Exception + public HARMIResponse invoke (long clientViewId, MarshalledObject mimo) throws Exception { RemoteMethodInvocation rmi = (RemoteMethodInvocation)mimo.get (); rmi.setMethodMap (invokerMap); @@ -133,10 +127,11 @@ try { HARMIResponse rsp = new HARMIResponse (); - if (tag lastSet) + long clusterViewId = partition.getCurrentViewId(); + if (clientViewId != clusterViewId) { rsp.newReplicants = new ArrayList (replicants); -rsp.tag = lastSet; +rsp.currentViewId = clusterViewId; } rsp.response = method.invoke (handler, rmi.getArguments ()); ___ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmx/src/main/org/jboss/ha/framework/interfaces HARMIResponse.java
User: vlada Date: 01/10/01 21:06:18 Modified:src/main/org/jboss/ha/framework/interfaces HARMIResponse.java Log: rather than using machine dependent system time to track if client's view of cluster has changed use logical viewID from cluster Revision ChangesPath 1.2 +1 -1 jbossmx/src/main/org/jboss/ha/framework/interfaces/HARMIResponse.java Index: HARMIResponse.java === RCS file: /cvsroot/jboss/jbossmx/src/main/org/jboss/ha/framework/interfaces/HARMIResponse.java,v retrieving revision 1.1 retrieving revision 1.2 diff -u -r1.1 -r1.2 --- HARMIResponse.java2001/09/30 15:19:32 1.1 +++ HARMIResponse.java2001/10/02 04:06:18 1.2 @@ -8,7 +8,7 @@ public ArrayList newReplicants = null; - public long tag = 0; + public long currentViewId = 0; public Object response = null; ___ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] tracking client's view of cluster using viewID
Hey Bill Sacha, As per our earlier conversation, I made this change in RH. I ran unit test and could not notice any unexpected behaviour. Vladimir Date: Mon, 1 Oct 2001 15:34:08 -0400 From: Bill Burke [EMAIL PROTECTED] To: Vladimir Blagojevic [EMAIL PROTECTED] Cc: Sacha Labourey [EMAIL PROTECTED] Subject: RE: Feedback Yes this is much better. Thanks. Keep it coming... Bill -Original Message- From: Vladimir Blagojevic [mailto:[EMAIL PROTECTED]] Sent: Monday, October 01, 2001 3:02 PM To: Bill Burke Cc: Sacha Labourey Subject: RE: Feedback Hey, viewId, is this a JG thing? Yes, that would probably be better. Yeah it is from JG. All members are in virtual synchrony, all see same views hence viewID is same across members until it changes by join/crash etc. Quick check on API . ViewId as object is retrievable from JChannel which is available from HARMIPartition. I'd use only public long ViewId.getId() to get long value viewId rather than whole object ViewId. Stuff this long value in HARMClient and upon replicant update check if clients viewid long is equal to the one we have on server , if not update client's recipient stubs. Seems ok? Vladimir ___ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] EXTENDED JMX 1.0 RI PACKAGE: NEW CONTRIBS!
Posted to jmx-forum by Christophe today... On 2001.06.26 13:51 Christophe Ebro wrote: Hello I am proud to announce that Sun has decided to provide in the packages of the JMX 1.0 RI some new contribs. These extensions are: Remoting - RemoteMBeanServer class allows on the client side to create an object representing a remote MBean Server, presenting the MBean Server interface and abstracting the underlying protocol. This object can then be directly called to have a single access point to the server side. - proxy interfaces a proxy is an object on the client side representing a remote MBean. Its abstracts the underlying protocol used to communicate with the remote MBean Server where the MBean lives. - RMI Adaptor Adaptor allowing a client to communicate with a remote MBean Server using RMI. It has both a client and a server parts: - the client part is used on the client side by the RemoteMBeanServer class described above. - the server part is instantiated on the agent side. MBeanServerForwarder Allows to customize the behavior of the default MBean Server, by implementing stackable MBean Servers: each level can handle one or more functionalities of the MBean Server, delegating the others to the underlying level. Context Allows to pass additional data (for example for security, access control) from the client to the MBean Server in the various requests. Such additional data can be handled at the server side by creating a MBeanServerForwarder, working with a MBeanServerChecker to check the context data. Heartbeat This mechanism is designed: - for a client to know if an agent is still alive (the client receives notifications that can be handled in user code) - for an agent to know if a client is still alive (does internal cleanup if client dies) CIM-JMX mapping and mof2MBean Tool to generate skeletons of standard MBean classes from CIM MOF schema(s). In the skeletons, code is generated to represent the CIM semantic (for example to check minimum and maximum values). The MBeans can then be instantiated and managed using any protocol (RMI, HTTP, etc.). It allows to: - model using CIM info model - generate MBeans - deploy them into a MBean Server - manage them using any protocol supported by the MBean Server Binaries, javadoc and examples are provided for those extensions. Note that the JMX 1.0 RI itself is unchanged. The HTML adaptor has been updated, to include latest fixes. The packaging has been slightly changed: the structure of the zip is: for the source package: jmx-1_0_1-ri/ contrib/- the new contribs jmx/ - the RI including the HTML adapter (like the previous package contents) for the binary package: jmx-1_0_1-ri_bin/ contrib/ jmx/ Have fun! chris/ Christophe Ebro - Sun Microsystems Java Technology For Service Providers - JMX Spec and Maintenance Leader OSS through Java initiative participant ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] changelog 2.4 beta
here is the text version (jsp is in cvs)if the release coordinator wants to include it in binary releases. [ #415088 ] Display LinkRef targets in JNDIView Added support for displaying the node that LinkRef JNDI nodes point to. This allows one to see what elements like ejb-refs, resource-refs, etc. in the ENC point to. [ #415379 ] Patch #415337: Informix jaws.xml See http://sourceforge.net/tracker/index.php?func=detailaid=415337group_id=22866atid=376687 [ #415982 ] ConfigurationService ServiceControl ConfigurationService ServiceControl The following changes have been made to the JBoss ConfigurationService ServiceControl mbeans: 1. The org.jboss.util.ServiceControl mbean service no longer listens for mbean registration events as the means for selecting mbeans which will receive the init/start/stop/destroy lifecycle method invocations. Mbeans wishing this service must register the org.jboss.util.Service interface though which they wish to be notified using the ServiceControl.register(Service) method. 2. The org.jboss.configuration.ConfigurationService mbean now registers all mbeans loaded from the jboss.jcml configuration file with the ServiceControl mbean. There is no requirement that an mbean implements the org.jboss.util.Service interface. The Service interface used to register the jboss.jcml mbean with the ServiceControl mbean is obtained either from the implementation of the new org.jboss.util.ServiceFactory interface specificed via the new serviceFactory jboss.jcml mbean tag attribute, or by a dynamic proxy that determines which if any of the init/start/stop/destroy methods the mbean implements using introspection of the mbean operations. The latter is used in the absence of a non-empty serviceFactory attribute. 3. Another behavior change made to the ConfigurationService is that it no longer reads the jboss-auto.jcml config file. This is now a write only file that does reflect configuration changes made at runtime, but it is no longer read on startup. [ #416083 ] Patch #415652: Jaws Mapping for SQL2000 See http://sourceforge.net/tracker/?func=detailatid=376687aid=415652group_id=22866 [ #416227 ] Resource adapters can be auto-deployed Resource adapters can be auto-deployed It is possible to configure the AutoDeployer service to use the RARDeployer service from the JBossCX module to deploy resource adapters. An example of this is included in jboss.jcml. This example sets up the directory deploy/lib to be watched for resource adapters. The bundled Minerva JDBC resource adapters are installed in this directory and are thus available at server startup. [ #416846 ] OpenJMS support migrated to contrib Support for OpenJMS as a JMS provider inside of JBoss has moved to an external contrib module: contrib/openjms The support classes have been repackaged into a single package in the contrib module, so users will need to update there configurations with the new package names: org.jboss.jms.openjms.OpenJMSProvider org.jboss.jms.openjms.OpenJMSServerSessionPoolFactory [ #416922 ] Don't
Re: [JBoss-dev] RMI-IIOP code scan
On 2001.06.18 14:48 Tahir Awan wrote: So, please help me with any tips to attack the code from the right angle. Just want to save some time wandering around and thinking if I got x hours to contribute, these should be utilized properly. Might want to look into somewhat outdated chapter 8 of online doco. Should help. Vladimir ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] Release notes summary for 2.2.1
Juha Scott, Here is the summary we talked about. If somebody has ideas were to put these , how to format and itergrate them etc etc speak out. 417433 LoginModule bug fixes Bug fix #415511 - JaasServerLoginModule is broken The 2.2 release of the JaasServerLoginModule is assigning all the union of all roles to each user. This was not seen in testing because the JBossSX unit- tests are not convering this login module. Bug fix #415818 - Shared principal name is not consistent The javax.security.auth.login.name value for the shared principal is not consistent with regard to type. Some modules are using a Principal, others the String name. At this level there is no reason not to consistently use the name of the principal as a String 417401 Added bundle build to contrib/tomcat Support for building bundled JBoss/Tomcat distributions was added to the contrib/tomcat cvs module. See the bundle target in the src/build/build.xml Ant buildfile. 416575 Fixed race in TxCapsule Bug fix #416377 It is possible under load for a TxCapsule to be released for reuse and for it to be picked for reuse before a 'commit' operation is actually complete. This _should_ only cause a problem if the transaction should have thrown a Hueristic exception of some sort, but it also causes lock contention, slowing one of the requests down a bit. 416920 Don't use CLASSPATH in run.bat/run.sh Removed the use of the system CLASSPATH variable as this results in too many conflicts in general. Changed CLASSPATH to JBOSS_CLASSPATH to make it clear that the CLASSPATH variable is not used. If you are relying on the system CLASSPATH to include required jars you either need to edit run.bat/run.sh to include CLASSPATH or setup a JBOSS_CLASSPATH variable with the required jars. 416438 Threading in HelloEJB Bug fix #416376 The HelloEJB servlet (for tomcat integration test) was not threadsafe. 415366 MetricsInterceptor improvement JMS Messages were never deliverd due to too small time to live value. 416812 Removed res-jndi-name java: prefix need Bug fix #414854 The requirement that the res-jndi-name be prefixed with java: has been removed from the ApplicationMetaData class. 417126 Updated InstantDB to 3.26 Updated the InstantDB idb.jar to version 3.26 to fix a deadlock issue. 417239 Fixed JDBCDataSourceLoader JDBCDataSourceLoader now works. It can be used to obtain a non-transactional JDBC connection pool. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] Rule Engine Integration Available
Hey, I saw your message too. However, I am so busy this week that I can not devote any time to JBoss. As I understand Rule engines are some sort of interpreter for if-then-else logic. I assume that they are going to be used mostly in business logic of your beans. I assume that instead of some if-then-else spaghetti , rule engine intercepts state mutations of an object , checks the rules against it and fires the rule event. Is this right? Some good intro I found here: www.haley.com Also there is jsr 94 about rule engine. Vlada On 2001.04.18 19:54 David Jencks wrote: Hi, At least one person saw my message. Thats one of the questions I was asking. A couple of months ago Marc suggested someone write a rule engine integration for jboss, so I volunteered. Although there is a java community process group working on an integration standard, I couldn't find out what they're up to, and I've had little luck finding out what kind of rules people might want to use. Any references or information or examples would be greatly appreciated. Thanks David Jencks On 2001.04.18 16:27:50 -0400 Jason Dillon wrote: What is a rule engine used for in real world applications? Could you give an example (other than "compute factorials and fibonacci numbers")? --jason On Wed, 18 Apr 2001, David Jencks wrote: Hi, I am pleased to announce that I have a working rule engine integration between JBoss and JEOPS (http://sourceforge.net/projects/jeops/). I have made the source available at http://home.earthlink.net/~davidjencks/. There are tar.gz and zip versions. Please follow the instructions in the README. You may need to change executable permissions on some scripts to get the examples to run. I would greatly appreciate all feedback positive and negative, and examples of how a rule engine could be useful. The examples I have so far are derived from the jeops examples, and are a bit artificial for ejb: they compute factorials and fibonacci numbers. They do serve as an example of essentially computing a transitive closure, not possible with simple queries. This implementation is similar to database triggers, in that rule engine interaction is not coded explicitly but triggered by calling an ejb method: an interceptor detects the call and passes it to the rule engine. The implementation is fully transactional: all work (including creating new beans, etc) is done within the original transaction provided by the original method call. ( A REQUIRES_NEW call in the middle of rule engine work will get a separate rule engine instance, not interfering with the original transaction). Jeops is somewhat unusual among rule engines in that rather than operating on tuples of values and interpreting rules, it operates on java objects (in this case remote interfaces) and compiles the rules into a java class. The main problem I am aware of at this time is that the implementation is extremely slow. This is only partly due to the very large amount of logging information generated. I have not yet investigated how to speed things up. Please, try it out, tell me how it works, and what more is needed. Thanks David Jencks ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] CVS update: newsite/business contrib.html binary.html cvs.html doco.html faq.html jboss-overview.html jboss-projects.html jboss-server.html jboss-third.html lists.html menu.html news.html team.html testimonials.html
On 2001.04.13 18:54 marc fleury wrote: | |At 18:21 13.4.2001 -0400, you wrote: |time to rewrite this site using taglibs don't you think ??? | |yes, maintaining the current site sucks big donkey dicks tell me about it... slurp slurp... Maybe it wouldn't be bad to build it from xml, as we do doco. http://www.nwalsh.com/website/index.html Vlada marc ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] DependencyManager jboss.dependencies
I think that the DependencyManager is a good one, so I would not be so quick to suggest that it be removed. Oh you dont want to start a war againsee dev archives. Yes it is not bad but I would say it is too explicit and verbose and doesn't do runtime dependencies. It would be nice to try to develop some solution using relations from jmx but it is currently not possible, it will be soon. http://archives.java.sun.com/cgi-bin/wa?A1=ind0104L=jmx-forum Vladimir --jason On Fri, 13 Apr 2001, marc fleury wrote: why don't you figure it out? let us know if it is depracated... marc |-Original Message- |From: [EMAIL PROTECTED] |[mailto:[EMAIL PROTECTED]]On Behalf Of Jason |Dillon |Sent: Friday, April 13, 2001 7:24 PM |To: [EMAIL PROTECTED] |Subject: [JBoss-dev] DependencyManager jboss.dependencies | | |Is this used anywhere? I could not find anything that read from |'jboss.dependencies' and it looks like the only place where the class is |even imported is in org.jboss.Main (though it is not instantiated). | |Just curious if this is just legacy or still needs to be hooked up. | |--jason | | |___ |Jboss-development mailing list |[EMAIL PROTECTED] |http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] JdbcProvider removal
Hi, Maybe it is not the best idea to change configuration(s) so often in between two minor releases. I would 1) tag removal of JdbcProvider scheduled for say 2.5 release. 2) support new way of specifying driver as attribute of DataSourceLoader 3) deprecate JDBCProvider , log that, but still allow user to use it until say 2.5 Just little a bit of buffer, makes us look more professional. Who knows Vladimir Juha-P Lindfors wrote: On Thu, 5 Apr 2001, Bill Burke wrote: Still want this done? If so, should I create a task(or Bug?) in SourceForge and assign it to myself? you should be able to add a todo task for yourself, still testing, let me know how it goes. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] Updated tomcat -jboss integration manual
Ken Jenks wrote: Vlada wrote: Hey, As somebody pointed out over the weekend - this doco was out of date in new manual. I would like to use this opportunity to ask maintainers of Tomcat-JBoss integration not to use old html doco but rather docbook. And someone should update the DOCUMENTATION-STANDARDS.html file that tells documentation developers to use the old HTML doco. Hey, +1 -- Ken Jenks, http://abiblion.com/ Tools for reading. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] Updated tomcat -jboss integration manual
Hey, As somebody pointed out over the weekend - this doco was out of date in new manual. I would like to use this opportunity to ask maintainers of Tomcat-JBoss integration not to use old html doco but rather docbook. Vlada ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[Fwd: [JBoss-dev] Documentation Confusion]
Forgot to cc jboss-dev. danch wrote: On the 'Documentation' page of the JBoss site, there's a link to the new manual. Below that is a note: The following content is already included in the new JBoss manual mentioned above. For the time of the ongoing converting process these documents will stay online. However, the 'Tomcat' section (for example) in the new manual still refers to the 2.0 Tomcat integration setup, whereas the 'howto' link Authors of Tomcat integration did not work on move to new doco. Tobias and I called , on several ocassions, people not to work on old doco but trather in Docbook. I'll devote my time and I'll translate it again (Tomcat) to Docbook tommorrow manual? The DB configuration section of the manual is primarily about 2.0, with a couple paragraphs at the end about 2.1. Can we make the 'new' manual about 2.1/2.2? DB stuff is updated 15 min ago, as well as configuration. Scott is currently updating more stuff. Actually, if I may say, I am really happy with the new documentation progress and I am sure that people who wrote in Docbook can confirm this. It is a real example of eXtreme Documentation at work !!! plgc and Doco Cheers, Vladimir thanks, danch ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] Sourceforge cvs problem
Hey Michael, Thanks a lot. I got it working (password-less ssh authentication) on RH7.0. I did not touch anything since I played with it last nite. So it was most likely this 6 hour propagation time you talked about. Thanks, Vladimir Micheal J wrote: Some ideas: 1. Use an RSA key-pair for SSH1. Ensure that the keys have no passphrase (no pint in going password-less only to be asked for the passphrase instead) 2. Use the "Edit key" feature of the account management page to upload your *public* key file only to all the relevant SF servers (this requires a 6 hour wait for full propagation) 3. Ensure your machine has the corresponding private+public keyfiles as your default (or only) SSH1 identity. Actually it may be enough for it to be just one of many identities but no point in complicating things right? Good luck, Micheal ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] Access
Filip Hanik wrote: the way to do this today is 1. get the list of all bound names 2. retrieve all the object bound from the JNDI tree 3. type check the returned object to see if it implements the EJBHome interface Am I correct? I would say so. 1. list bindings in root context of global namespace (all deployed homes are found here) 2. select only those who are (Proxy.isProxyClass(object.getClass()) and (object instanceof EJBHome))) (although the first check is redundant, just for fun) Vladimir ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] Sourceforge cvs problem
Hey Fulco, Almost same config as yours. You can not use pserver authentication to commit, you have to use ssh. Get an ssh client for linux from gnomerpm if you dont have it. Do a clean checkout to get rid of CVS directories that contain old method of authentication in Root file. Do co without login command and then you can commit. Note that you still have to authenticate every time you issue cvs command. I tried to upload keys to prevent this but without success. If someone has done this , please respond. Cheers, Vladimir Fulco Muriglio wrote hallo, I'm not behind a firewall and I use Linux, cvs works except for commit: the response is: cvs [server aborted]: "commit" requires write access to the repository Is it a problem of my account? The command was: cvs -z3 -d:pserver:[EMAIL PROTECTED]:/cvsroot/jboss commit contrib/catalina but I've tried also: cvs -z3 -d:pserver:[EMAIL PROTECTED]:/cvsroot/jboss commit contrib/catalina With the same result. Bye Fulco ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] Sourceforge cvs problem
Hey, I still can not do this on Redhat 7 although I am close. Used ssh-keygen, created both RSA and DSA key pairs, added them to running ssh-agent, uploaded pub keys to server. Did you use website upload or scp command? In the log below clearly server refuses my private RSA key . ssh -v -l vlada cvs.jboss.sourceforge.net OpenSSH_2.5.2p2, SSH protocols 1.5/2.0, OpenSSL 0x0090581f debug1: Seeding random number generator debug1: Rhosts Authentication disabled, originating port will not be trusted. debug1: ssh_connect: getuid 500 geteuid 0 anon 1 debug1: Connecting to cvs.jboss.sourceforge.net [216.136.171.202] port 22. debug1: Connection established. debug1: identity file /home/vlad/.ssh/identity type 0 debug1: unknown identity file /home/vlad/.ssh/id_rsa debug1: identity file /home/vlad/.ssh/id_rsa type -1 debug1: identity file /home/vlad/.ssh/id_dsa type 2 debug1: Remote protocol version 1.99, remote software version OpenSSH_2.5.1p1 debug1: match: OpenSSH_2.5.1p1 pat ^OpenSSH debug1: Local version string SSH-1.5-OpenSSH_2.5.2p2 debug1: Waiting for server public key. debug1: Received server public key (768 bits) and host key (1024 bits). debug1: Host 'cvs.jboss.sourceforge.net' is known and matches the RSA1 host key. debug1: Found key in /home/vlad/.ssh/known_hosts:2 debug1: Encryption type: 3des debug1: Sent encrypted session key. debug1: Installing crc compensation attack detector. debug1: Received encrypted confirmation. debug1: Trying RSA authentication via agent with '[EMAIL PROTECTED]' debug1: Server refused our key. debug1: RSA authentication using agent refused. debug1: Trying RSA authentication with key '[EMAIL PROTECTED]' debug1: Server refused our key. debug1: Doing password authentication. [EMAIL PROTECTED]'s password: Scott M Stark wrote: I have been able to upload my ssh key to allow for access without having to type the password each time. I used ssh-keygen and did not assign the generated key a passphrase. This is only working on windows 2000 though. On RedHat linux 7.0 I am having to enter the password each time. I believe this is an issue with the linux ssh using version 2 of the protocol vs version 1.5 on windows but I have not looked in the issue in detail. You can debug the secure connection setup by issuing the following command: ssh -v -l userid cvs.jboss.sourceforge.net cvs -t server You can add more than one -v arg to make the ssh output increasingly verbose. The -t to the cvs command puts it in trace mode so that it also tells you what its doing. I tried to upload keys to prevent this but without success. If someone has done this , please respond. Cheers, Vladimir ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development