Re: JDK requirement
I don't think the requirement is on the jdk but on the jre. On Thursday, August 5, 2010, S Brady step...@bitlev.com wrote: I am new to Karaf and OSGi in general, but I'm trying to understand why Karaf has a JDK requirement if it is a runtime container instead of just JRE. Is there a way I can work around this requirement? If so, how? Or what assumption/context am I not getting? -- View this message in context: http://karaf.922171.n3.nabble.com/JDK-requirement-tp1027660p1027660.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Karaf, Camel and startup race conditons
The issue comes from Spring-DM which has a big flaw in the sense it can not provide any way to handle those dependencies. We had the same problem and the best way to solve those is to use Aries Blueprint and Camel Blueprint support instead of spring. You could try playing with the bundle start level, but I won't solve all the problems as spring does some things asynchronously wrt the bundles starting, so you'll still have race conditions. It might help a bit though, so feel free to give it a try. I think the easiest way would be to configure fileinstall so that the bundles deployed have a bundle start level greater than the default bundle start level. I can't lookup the exact mechanism right now, but will try this evening if no one gave it to you (I think Charles has experimented things like that IIRC). On Mon, Aug 16, 2010 at 06:21, Thiago Souza tcostaso...@gmail.com wrote: Hello there, I've setup a production server based on karaf and camel. I've installed camel through the features mechanism and deployed several spring xml files that defines components and camel contexts instantiation. Everything is fine until the server is restarted. In this case I get all sort of random problem like type converters that can not be found, routes based on quartz throwing exceptions and other weird behavior. The only way to solve this problem is to undeploy all the xml files, restart the system and redeploy them in a controlled way. How can I solve this? Is there any way to start the xmls AFTER all the camel bundles and have them started in an ordered way? Regards, Thiago Souza -- View this message in context: http://karaf.922171.n3.nabble.com/Karaf-Camel-and-startup-race-conditons-tp1165020p1165020.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Setting karaf.instances system property has no effect
I think it would be better if you raised the issue yourself. Anyone can create a JIRA issue, you just need to create an account on JIRA. It seems you also have a patch ready, so if you could attach it, that would be even better :-) On Wed, Aug 18, 2010 at 16:27, odeak o...@ztec.ch wrote: Hi I'm trying to separate the static files from the dynamic ones in a karaf deployment but setting the karaf.instances system property is ignored (karaf 2.0.0). This is because in Main.java on line 191, validate=false when calling Utils.getKarafDirectory(...) so it always revert to the default. Could you please raise an issue to fix this? Regards, Oliver -- View this message in context: http://karaf.922171.n3.nabble.com/Setting-karaf-instances-system-property-has-no-effect-tp1206801p1206801.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Strange behaviour with SSH
Not sure what's wrong as the parameters seem correct:
http://svn.apache.org/repos/asf/karaf/tags/karaf-2.0.0/shell/wrapper/src/main/resources/org/apache/karaf/shell/wrapper/unix/karaf-wrapper.conf
You could try to turn on debug logging for JSW and see if the actual
parameters are correct.
The config file should be located in the etc/ folder after running the
wrapper:install command.
On Tue, Aug 24, 2010 at 10:24, Achim Nierbeck achim.nierb...@ptv.de wrote:
Sorry, forgot to tell you that I'm using the 2.0 version of the Karaf server.
The jce.jar is included in the lib folder ${JAVA_HOME}/jre/lib of the jdk
(was downloaded last week from the dl-server)
Any Idea how to configure JSW (JavaServiceWrapper?) different.
Guillaume Nodet wrote:
This problem is usually caused by the JCE libraries not being included
in the classpath.
On some systems, the jars for cryptography are located in
${JAVA_HOME}/jre/lib/endorsed or ${JAVA_HOME}/lib/endorsed. Those
folders are configured in the batch file, but maybe there's something
wrong in the way the JSW is configured. Does that happen using trunk
/ 2.0.0 too or only 1.6.0 ?
On Tue, Aug 24, 2010 at 10:08, Achim Nierbeck achim.nierb...@ptv.de
wrote:
Hi,
I do have a strange behavior when trying to connect to the Karaf server
on a
Linux machine with ssh.
If I do run the Karaf as a shell application (starting through the Karaf
shell script) I'm able to connect to it via SSH. If I run the Karaf as a
service with the service wrapper I'm not able to connect to it, which is
quite strange.
Googling for this strange behavior showed me that there was some sort of
issue with the SSH implementation back in 1.60 which should be fixed by
now.
I do get the following exception:
[ INFO] 10:00:04,066 (NioProcessor-1) Session created...
[ INFO] 10:00:04,086 (NioProcessor-1) Client version string:
SSH-2.0-PuTTY_Release_0.60
[ INFO] 10:00:04,086 (NioProcessor-1) Received SSH_MSG_KEXINIT
[ WARN] 10:00:04,088 (NioProcessor-1) Exception caught
java.lang.IllegalStateException: Unable to negociate key exchange for
item 2
at
org.apache.sshd.common.session.AbstractSession.negociate(AbstractSession.java:886)
at
org.apache.sshd.server.session.ServerSession.handleMessage(ServerSession.java:151)
at
org.apache.sshd.common.session.AbstractSession.decode(AbstractSession.java:522)
at
org.apache.sshd.common.session.AbstractSession.messageReceived(AbstractSession.java:225)
at
org.apache.sshd.common.AbstractSessionIoHandler.messageReceived(AbstractSessionIoHandler.java:58)
at
org.apache.mina.core.filterchain.DefaultIoFilterChain$TailFilter.messageReceived(DefaultIoFilterChain.java:713)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.filterchain.DefaultIoFilterChain.callNextMessageReceived(DefaultIoFilterChain.java:434)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.filterchain.DefaultIoFilterChain.access$1200(DefaultIoFilterChain.java:46)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.filterchain.DefaultIoFilterChain$EntryImpl$1.messageReceived(DefaultIoFilterChain.java:793)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.filterchain.IoFilterAdapter.messageReceived(IoFilterAdapter.java:119)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.filterchain.DefaultIoFilterChain.callNextMessageReceived(DefaultIoFilterChain.java:434)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.filterchain.DefaultIoFilterChain.fireMessageReceived(DefaultIoFilterChain.java:426)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.polling.AbstractPollingIoProcessor.read(AbstractPollingIoProcessor.java:638)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.polling.AbstractPollingIoProcessor.process(AbstractPollingIoProcessor.java:598)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.polling.AbstractPollingIoProcessor.process(AbstractPollingIoProcessor.java:587)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.polling.AbstractPollingIoProcessor.access$400(AbstractPollingIoProcessor.java:61)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.core.polling.AbstractPollingIoProcessor$Processor.run(AbstractPollingIoProcessor.java:969)[31:org.apache.mina.core:2.0.0.RC1]
at
org.apache.mina.util.NamePreservingRunnable.run(NamePreservingRunnable.java:64)[31:org.apache.mina.core:2.0.0.RC1]
at
java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)[:1.6.0_21]
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)[:1.6.0_21]
at java.lang.Thread.run(Thread.java:619)[:1.6.0_21]
[ INFO] 10:00:04,092 (NioProcessor-1) Closing session
--
View this message in context:
http://karaf.922171.n3.nabble.com/Strange-behaviour-with-SSH-tp1306643p1306643.html
Sent
Re: Karaf 2.0.0 Jersey 1.3.0
JB, you're right, but it still should not cause a resolution failure, especially if a single version of this bundle is installed in the framework. Jérôme, I would try to switch to equinox (in the etc/config file) and restart to see what happen. Also, the Felix Framework 3.0.2 has been released recently which may fix the problem, so it's worth a try too. I'll upgrade to this latest version in trunk asap anyway. On Thu, Aug 26, 2010 at 10:20, Jean-Baptiste Onofré j...@nanthrax.net wrote: Hi Jérôme, The com.sun.jersey.api package is provided by jersey-core jar. It's exported by the jersey-core bundle (according to the MANIFEST). You are right, the package shouldn't be imported and exported. The bundle package statement should looks like : Export-Package com.sun.jersey*... /Export-Package Import-Package !com.sun.jersey*, * /Import-Package Like this, it can support several version of jersey-core bundle in the same OSGi framework and avoid this kind of issue. A possible workaround is to create your own jersey-core bundle using maven-bundle-plugin for example. Regards JB On 08/26/2010 10:10 AM, Jérôme Pochat wrote: Hi I tried to install Jersey REST implementation into fresh Karaf 2.0.0 instance without any success. This was working fine in previous version of Karaf. What is strange is that there is no problem to install it inside Felix 3.0.*. install jsr311-api-1.1.1.jar == Success install jersey-core-1.3.jar == Failed Error executing command: Unresolved constraint in bundle com.sun.jersey.jersey-core [34]: Unable to resolve 34.0: missing requirement [34.0] package; (package=com.sun.jersey.api.representation) - [34.0] package; (package=com.sun.jersey.api.representation) This package is imported and exported by the same bundle. Could it be a reason of the issue? Thanks in advance for help. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Please enlighten me
If a bundle explicitely import the package, I don't think boot delegation will be enough, as the resolver will need someone to export the package nonetheless (afaik). On Fri, Aug 27, 2010 at 12:53, Achim Nierbeck achim.nierb...@ptv.de wrote: OK, this is exactly the way I expected it to be. But where am I wrong in expecting that the package com.sun.org.apache.xerces.internal.dom is already available through the bootdelegation. Why did I need to configure org.osgi.framework.system.packages.extra or I could probably also have configured org.osgi.framework.system.packages Or is it just that I need to tell the system.packages through either org.osgi.framework.system.packages or org.osgi.framework.system.packages.extra and specify through the org.osgi.framework.bootdelegation that certain packages are only available through the bootdelegation? Till now I thought that this is implicitly done through the bootdelegation parameter. Guillaume Nodet wrote: org.osgi.framework.system.packages.extra = Framework environment property identifying extra packages which the system bundle must export from the current execution environment. This property is useful for configuring extra system packages in addition to the system packages calculated by the framework. org.osgi.framework.bootdelegation = Framework environment property identifying packages for which the Framework must delegate class loading to the parent class loader of the bun- dle. Boot delegation does not require a bundle to import the pacakge, whereas the extra property will only make the system bundle export the given packages. On Fri, Aug 27, 2010 at 11:25, Achim Nierbeck achim.nierb...@ptv.de wrote: How does the property org.osgi.framework.system.packages.extra work different than the org.osgi.framework.bootdelegation Because com.sun.* is set in the bootdelegation but the bundle Apache ServiceMix Bundles: saaj-impl-1.3.2 (1.3.2.1) wasn't able to get access to com.sun.org.apache.xerces.internal.dom when configuring the org.osgi.framework.system.packages.extra with this package (com.sun.org.apache.xerces.internal.dom) it was starting. As far as I understood the bootdelegation should already have done this, right? BTW, using Karaf 2.0.0 and I found the hint by accident on Jamie Goodyears Blog Thanx, Achim -- View this message in context: http://karaf.922171.n3.nabble.com/Please-enlighten-me-tp1367765p1367765.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- View this message in context: http://karaf.922171.n3.nabble.com/Please-enlighten-me-tp1367765p1368720.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Karaf 2.0.0 Jersey 1.3.0
Yeah, Karaf is consistent so the same packages are exported whtether you use felix or equinox (or it should be). However, we have dicussed not restricting the packages provided by the jre, so i think we need to get back to this discussion. On Fri, Sep 3, 2010 at 17:19, Jérôme Pochat jpoc...@sopragroup.com wrote: Hi again I'm back :) Since I switch to Equinox as workaround, I now face new issue about Java 6 API that is also provided by ServiceMix bundles (i.e. StAX). Using Felix, some system packages are hidden in jre.properties. Using Equinox, all system packages seems to be exported. In some cases (such as mine :-p), this could create conflicts at runtime (java.lang.LinkageError). Does jre.properties file is read with using Equinox as well as Felix? Thanks in advance. -- View this message in context: http://karaf.922171.n3.nabble.com/Karaf-2-0-0-Jersey-1-3-0-tp1347655p1412755.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Karaf 2.0.1-SNAPSHOT and webconsole, cannot login
Jean-Baptiste is working on the JAAS side to add new features (the ability to have encrypted passwords for better security). The problem was a small regression caused by the ongoing work, that's all On Mon, Sep 13, 2010 at 08:56, Bengt Rodehav be...@rodehav.com wrote: Hi Guillaume, I wasn't sure that it was an issue. Was it a missing JAAS configuration? /Bengt 2010/9/13 Guillaume Nodet gno...@gmail.com I've just fixed the issue. Thx for reporting! On Sun, Sep 12, 2010 at 22:18, Bengt Rodehav be...@rodehav.com wrote: I looked at the log a bit more carefully and found: javax.security.auth.login.LoginException: Inga inloggningsmoduler har konfigurerats för karaf The above has been translated to Swedish but in English it says that no login modules have been configured for karaf. I know I should probably know how to do this but if anyone can give me some assistance I would really appreciate it. I also think that it is a good idea to have an easy login configured from scratch when installing Karaf - it's a lot easier to customize an existing configuration than creating a configuration. /Bengt 2010/9/12 Bengt Rodehav be...@rodehav.com I've installed Karaf 2.0.1-SNAPSHOT and installed the Webconsole feature. However, I cannot login to the web console. I've always logged in (on Karaf 1.6.0 and also Karaf 2.0.0) with user=karaf and pwd=karaf. If I enter those credentials the login dialog just pops up again, and again, and again... I've found someone with similar problems that recommended that a file with the name org.apache.karaf.webconsole.cfg should be placed in the etc directory. The contents should be: config name=org.apache.karaf.webconsole realm=karaf /config I've tried that but the problem remains. Can anyone help me out here? /Bengt -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Karaf - Certificate Keystore
I suppose you're talking about connecting through the SSH console, right ? In that case, the answer is no, because it can't be done atm. The SSH layer does not configure any PublicKeyAuthenticator so that you can't rely on SSH key based authentication. I do agree this is something we should do, it's just that I've never had the time to work on it. See KARAF-32, it's an old bug. Another related thing is captured in KARAF-32, which is the use of agent forwarding to be able to log into another karaf instance using the same credentials. Currently, you have to give them again, even if you're already authenticated, because they are not kept in memory. This would be much better. I guess one possible problem is the use of the local console. The user isn't really authenticated and we have no way to know the password or public key in that case. Not sure how to handle that yet. On Mon, Sep 27, 2010 at 17:21, Charles Moulliard cmoulli...@gmail.comwrote: Hi, Do we have an example or a test case showing how to configure karaf to use certificate + keystore to authenticate admin user instead of simple login mechanism provided by default ? Regards, Charles Moulliard Senior Solution Architect - Fuse Consultant Open Source Integration: http://fusesource.com Blog : http://cmoulliard.blogspot.com Twitter : http://twitter.com/cmoulliard Linkedin : http://www.linkedin.com/in/charlesmoulliard Skype: cmoulliard -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: features.xml - Option to dicate JDK levels
The nice thing I see with OBR is that the resolution is much closer to the real environement, as OBR takes into account the already deployed bundles. For example if you try to install camel-blueprint on Karaf 2.1, the whole environment is kinda messed up because you end up with two versions of the aries blueprint implementation (which does not work very well). Ideally, the camel-blueprint feature would put a requirement on a blueprint extender being present, and this requirement would be solved by the current environment, so obr would not try to install another version of blueprint. In our case, for camel-jaspyt, it 's not very easy to model (though it's possible), but I honestly thing this is a very rare case. The way to model that would be to have the camel-jasypt bundle (or is it the jasypt bundle itself?) have a requirement on a specific capability, let's say foo which happen to be provided by the iucla bundle and also by the system bundle (or the environment) if running on jdk6. Though, IIUC it does not really harm to install icu4j even on JDK 6, so at worst, there's an unused bundle in the system. On Tue, Oct 12, 2010 at 08:25, Jean-Baptiste Onofré j...@nanthrax.net wrote: Guillaume, Where to put the bundle resolution logic in the OBR ? The user should be able to define the bundle set, so maybe the easiest location is the feature descriptor. I know that the OBR supports properties on repo, but the user will have to tune the repo to add some properties for bundle provisioning. I wonder what the easiest way for the user to define it. Regards JB On 10/12/2010 08:10 AM, Guillaume Nodet wrote: I'm not sure we should add too much of this in the features descriptors. I think a better idea would be to start leveraging OBR to determine the best set of dependencies for a given set of bundles to install. If needed we could also leverage the obr url handler to use a filter to actually select a bundle. On Tuesday, October 12, 2010, Jean-Baptiste Onofréj...@nanthrax.net wrote: Hi Claus, Up to now, AFAIK, it's not possible to define a feature with JDK specific bundles (the descriptor is static). You can add some JRE/JDK specific definition in etc/jre.properties but it's global to the kernel (not dedicated to a given feature). Anyway, I think it's interesting. We can extend the feature deployer to support this kind of conditions. I'm gonna raise a Jira task around this. Regards JB On 10/12/2010 06:16 AM, Claus Ibsen wrote: Hi I wonder if its possible in the features.xml file to define a bundle being qualified depending on the current JDK? For example if you run JDK 1.5 you want the bundle included. If you run JDK 1.6+ you do NOT. The option should most likely support a range similar to the OSGi versioning. Maybe something similar to this: bundle jdk=[1.5,1.6)mvn:xxx/yyy/2.2/bundle An example would be many of the encryption frameworks which requires additional jars to run on JDK 1.5, where as 1.6 provides API and chipers out of the box. And we could have a similar situation when JDK 1.7 comes out. Where you may need additional JARs on 1.6 and not on 1.7. I could not find such information at http://karaf.apache.org/46-provisioning.html But it could be the documentation is outdated -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: How to avoid that log4j classes are loaded twice
That's where the JEE fun begins. I suppose you need to configure you JEE web server to not make the log4j classes available to the war. I don't think there's a standard way of doing that though ... Is the behavior the same wether you use Felix or Equinox ? On Thu, Oct 21, 2010 at 13:50, Charles Moulliard cmoulli...@gmail.comwrote: Hi, When Karaf is deployed as a WAR in Jetty or Tomcat, log4j classes are loaded twice and of course Karaf is not able to report trace in servicemix.log file. log4j:ERROR A org.apache.log4j.ConsoleAppender object is not assignable to a org.apache.log4j.Appender variable. log4j:ERROR The class org.apache.log4j.Appender was loaded by log4j:ERROR [4.0] whereas object of type log4j:ERROR org.apache.log4j.ConsoleAppender was loaded by [contextloa...@servicemix Embedded Example]. log4j:ERROR Could not instantiate appender named A1. log4j:ERROR A org.apache.log4j.ConsoleAppender object is not assignable to a org.apache.log4j.Appender variable. log4j:ERROR The class org.apache.log4j.Appender was loaded by log4j:ERROR [4.0] whereas object of type log4j:ERROR org.apache.log4j.ConsoleAppender was loaded by [contextloa...@servicemix Embedded Example]. Does anybody has an idea to avoid that ? Regards, Charles -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: How about to to facilitate to use JNDI(boot delegation) in karaf?
The boot delegation property can be modified in etc/config.properties or overriden in etc/custom.properties. Can you give a bit more detail on how to use JNDI ? I'm not sure to understand why the package is used but not directly referenced. On Fri, Nov 5, 2010 at 10:35, ext2 x...@tongtech.com wrote: Hi: When I using jndi in osgi, the javax.naming package is required. But because javax.naming package is not direct refereced in the java's source code(; So the generated bundle's Import-Package header doesn't contains javax.naming package; Although the karaf's etc/jre.properties file could define javax.naming to be exportable from boot class loader. But because we doesn't declare import for javax.naming package in bundle,so a ClassNotFound Exception will still occurs; So we want the package javax.naming which could be used, even there is not import declaration in bundle; To support this, I can only define org.osgi.framework.bootdelegation system property in karaf's startup script; So I am wondering if karaf could support a configure file which allow we configure the org.osgi.framework.bootdelegation ? if so , it's easy to use control boot delegation; -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Features in Karaf / Pax Exam
I think pax-exam has its own features parser, so it may have been kept
behing what karaf provides.
We really need to use a schema and version that to be able to detect
such problems more easily.
On Fri, Nov 5, 2010 at 10:17, Andreas Gies andr...@wayofquality.de wrote:
Hello,
I am sorry if am hitting the wrong list with my question, but it seemed a
good starting point for me ...
We have started the development of a karaf based application and wnat to
exploit the features support
for packaging. In my features I have the following:
features
feature name=YConX::Whiteboard version=${project.version}
!-- This line works for deploying into karaf, but not in conjunction with
pax - exam --
!--feature version=${project.version}YConX::Equinox/feature--
!-- TODO: avoid repetition of the derby version --
bundlemvn:org.apache.derby/derby/10.6.2.1/bundle
bundlemvn:${project.groupId}/${project.groupId}.core/${project.version}/bundle
!--bundlemvn:${project.groupId}/${project.groupId}.database/${project.version}/bundle
bundlemvn:${project.groupId}/${project.groupId}.dataproviderservice/${project.version}/bundle
bundlemvn:${project.groupId}/${project.groupId}.gui/${project.version}/bundle
bundlemvn:${project.groupId}/${project.groupId}.logger/${project.version}/bundle--
/feature
/features
For my integration test the first thing i want to test is that the packaging
works ok and have copied/modified
the FeaturesTest found in karaf itself. In my configuration I have:
@Configuration
public static Option[] configuration() throws Exception {
return OptionUtils.combine(
Helper.getDefaultOptions(
systemProperty(org.ops4j.pax.logging.DefaultServiceLog.level).value(DEBUG)
),
PaxRunnerOptions.scanFeatures(
maven().groupId(de.yconx.osgi).artifactId(features).type(xml).classifier(features).version(0.0.1-SNAPSHOT),
YConX::Equinox/0.0.1-SNAPSHOT
),
PaxRunnerOptions.scanFeatures(
maven().groupId(de.yconx.whiteboard).artifactId(features).type(xml).classifier(features).version(0.0.1-SNAPSHOT),
YConX::Whiteboard/0.0.1-SNAPSHOT
),
waitForFrameworkStartup(),
equinox()
);
}
The test succeeds if I omit the feature reference in my feature definition,
but fails if i include it.
The basic idea was, that the feature is my unit of deployment into the karaf
container and that features
kind of build on top of each other.
So, the question is:
Am I doing something out of the ordinary ?
Am I hitting a pax-exam limitation that is already known? (happy to log a
bug there...)
Is there a syntax issue I am not spotting and would make it work in both
environments ?
Thanks in advance
Andreas
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: Change ka...@root prompt?
The default PROMPT is defined as:
public static final String DEFAULT_PROMPT =
\u001b[1m${user}\u001b...@${application} ;
To change this, you can set the PROMPT variable in
etc/shell.init.script to something that suit your needs better.
Either as a plain variable:
PROMPT=\u001b[1m${user}\u001b...@${application} ;
Or as a closure
'#PROMPT' = { echo \u001b[1m${user}\u001b...@${application}ww }
The branding does not currently enable changing the default prompt,
but maybe it could be enhanced.
On Fri, Nov 5, 2010 at 19:29, Kit Plummer kitplum...@gmail.com wrote:
Is it possible to change the karaf part of the console prompt in
configuration?
Kit
--
View this message in context:
http://karaf.922171.n3.nabble.com/Change-karaf-root-prompt-tp184p184.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: Change ka...@root prompt?
You're right. The correct value for karaf 2.1.0 would be:
PROMPT= '\u001b\\[1m${user}\u001b\\...@${application} ';
Though the value I provided will work on 2.2.0 (trunk for now).
On Fri, Nov 5, 2010 at 21:15, Mike Van mvangeert...@comcast.net wrote:
Guillaume,
Are you sure that's correct? I added:
PROMPT=\u001b[1m${user}\u001b...@${application} ;
to my shell.init.script file, and I received the following error in the
console:
Error in initialization script: Eof found in the middle of a compound for
'][', begins at b...@${application}
v/r,
Mike Van
- Original Message -
From: Guillaume Nodet [via Karaf]
ml-node+1850211-615221409-228...@n3.nabble.com
To: Mike Van mvangeert...@comcast.net
Sent: Friday, November 5, 2010 3:07:07 PM
Subject: Re: Change ka...@root prompt?
The default PROMPT is defined as:
public static final String DEFAULT_PROMPT =
\u001b[1m${user}\u001b...@${application} ;
To change this, you can set the PROMPT variable in
etc/shell.init.script to something that suit your needs better.
Either as a plain variable:
PROMPT=\u001b[1m${user}\u001b...@${application} ;
Or as a closure
'#PROMPT' = { echo \u001b[1m${user}\u001b...@${application}ww }
The branding does not currently enable changing the default prompt,
but maybe it could be enhanced.
On Fri, Nov 5, 2010 at 19:29, Kit Plummer [hidden email] wrote:
Is it possible to change the karaf part of the console prompt in
configuration?
Kit
--
View this message in context:
http://karaf.922171.n3.nabble.com/Change-karaf-root-prompt-tp184p184.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
View message @
http://karaf.922171.n3.nabble.com/Change-karaf-root-prompt-tp184p1850211.html
To start a new topic under Karaf - User, email
ml-node+930749-917263437-228...@n3.nabble.com
To unsubscribe from Karaf - User, click here .
--
View this message in context:
http://karaf.922171.n3.nabble.com/Change-karaf-root-prompt-tp184p1850561.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: Bundle lost after shutdown
The problem happens because the features service does not remove the blank characters before and after the bundles urls and it looks like the felix framework will not reload bundles with such invalid urls. I've just raised KARAF-268 which I'll try to fix later today. On Mon, Nov 8, 2010 at 20:23, Jorge Riquelme to...@totex.cl wrote: Guillaume, I can reproduce the issue with karaf 2.1.0 and the attached files: a test feature file, plus custom.properties and jre.properties (stolen from smix :p). First, load the features.xml file and install the contentcompass feature: ka...@root features:addurl file:/home/totex/features.xml ka...@root features:install contentcompass ka...@root list START LEVEL 100 , List Threshold: 50 ID State Blueprint Level Name [ 31] [Active ] [Created ] [ 60] Apache Karaf :: Shell ConfigAdmin Commands (2.1.0) [ 32] [Active ] [ ] [ 60] Apache Aries Transaction Manager (0.2.0.incubating) [ 33] [Active ] [ ] [ 60] Apache ServiceMix :: Specs :: Stax API 1.0 (1.6.0.SNAPSHOT) ... [ 70] [Active ] [ ] [ 60] Apache CXF Bundle Jar (2.2.11) ... [ 84] [Active ] [ ] [ 60] Clerezza Ext - Jena OSGi Bundle (0.6.0.incubating-SNAPSHOT) All bundles are active. Then, shutdown (ctrl+d) and restart: ka...@root ERROR: Error starting mvn:org.apache.cxf/cxf-bundle/2.2.11 (org.osgi.framework.BundleException: Unresolved constraint in bundle org.apache.cxf.bundle [70]: Unable to resolve 70.0: missing requirement [70.0] package; (package=javax.transaction.xa)) org.osgi.framework.BundleException: Unresolved constraint in bundle org.apache.cxf.bundle [70]: Unable to resolve 70.0: missing requirement [70.0] package; (package=javax.transaction.xa) at org.apache.felix.framework.Felix.resolveBundle(Felix.java:3409) at org.apache.felix.framework.Felix.startBundle(Felix.java:1709) at org.apache.felix.framework.Felix.setActiveStartLevel(Felix.java:1143) at org.apache.felix.framework.StartLevelImpl.run(StartLevelImpl.java:264) at java.lang.Thread.run(Thread.java:636) ka...@root list START LEVEL 100 , List Threshold: 50 ID State Blueprint Level Name [ 31] [Active ] [Created ] [ 60] Apache Karaf :: Shell ConfigAdmin Commands (2.1.0) [ 33] [Active ] [ ] [ 60] Apache ServiceMix :: Specs :: Stax API 1.0 (1.6.0.SNAPSHOT) [ 34] [Active ] [ ] [ 60] Stax2 API (3.0.2) ... [ 70] [Installed ] [ ] [ 60] Apache CXF Bundle Jar (2.2.11) ... [ 84] [Active ] [ ] [ 60] Clerezza Ext - Jena OSGi Bundle (0.6.0.incubating-SNAPSHOT) Bundle 32 is lost, and bundle 70 isn't running for the missing dependency. Then: ka...@root install -s mvn:org.apache.aries.transaction/org.apache.aries.transaction.manager/0.2-incubating Bundle ID: 85 ka...@root restart 70 After manually reinstall the bundle, the problem doesn't happen anymore (all loads fine after shutdown and restart). saludos 2010/11/8 Guillaume Nodet gno...@gmail.com: Can you reproduce the problem easily ? If so, could you please give the exact steps you use to reproduce the problem ? On Mon, Nov 8, 2010 at 04:21, Jorge Riquelme to...@totex.cl wrote: Hi list, i'm having a problem with karaf 2.1.1-SNAPSHOT with a particular bundle (mvn:org.apache.aries.transaction/org.apache.aries.transaction.manager/0.2-incubating). I start from a fresh install of karaf and deploy my feature; all fine: ka...@root list START LEVEL 100 , List Threshold: 50 ID State Blueprint Spring Level Name ... [ 43] [Active ] [ ] [ ] [ 60] spring-osgi-extender (1.2.0) [ 44] [Active ] [ ] [ ] [ 60] spring-osgi-annotation (1.2.0) [ 45] [Active ] [ ] [ ] [ 60] Apache Aries Transaction Manager (0.2.0.incubating) [ 46] [Active ] [ ] [ ] [ 60] Apache ServiceMix :: Specs :: Stax API 1.0 (1.6.0.SNAPSHOT) [ 47] [Active ] [ ] [ ] [ 60] Stax2 API (3.0.2) ... After, when I restart karaf, the bundle 45 is lost and I get several exceptions from the other dependent bundles (of aries tx): ka...@root ERROR: Error starting mvn:org.apache.cxf/cxf-bundle/2.2.11 (org.osgi.framework.BundleException: Unresolved constraint in bundle org.apache.cxf.bundle [83]: Unable to resolve 83.0: missing requirement [83.0] package; (package=javax.transaction.xa)) org.osgi.framework.BundleException: Unresolved constraint in bundle org.apache.cxf.bundle [83]: Unable to resolve 83.0: missing requirement [83.0] package; (package=javax.transaction.xa) at org.apache.felix.framework.Felix.resolveBundle(Felix.java:3409) at org.apache.felix.framework.Felix.startBundle(Felix.java:1709) at org.apache.felix.framework.Felix.setActiveStartLevel(Felix.java:1143
Re: Open Ports in Karaf
One of them must be the tcp port opened to stop karaf as in tomcat (i.e. a free random port is opened and bound to the localhost only and written to the data/port file). It can be configured using the karaf.shutdown.port property (the default value of 0 means a random port). Not sure about the other one though. On Tue, Nov 9, 2010 at 17:56, Roshan A. Punnoose rpunno...@proteuseng.com wrote: Hi, I am trying to figure out what ports Karaf opens up. I was able to get most of them (ssh, jmx rmi, etc) except a few. Everytime Karaf starts up (v. 2.1.0) it opens up two randomly assigned ports in the high numbers. For example: 3:tcp4 0 0 127.0.0.1.57165 *.* LISTEN 4:tcp46 0 0 *.57164 *.* LISTEN If I turn off the DisableAttachMechanism flag so that jconsole cannot connect to karaf, the bottom one is gone, but the top one remains: 3:tcp4 0 0 127.0.0.1.57168 *.* LISTEN (Notice, they are different numbers because they are randomly assigned) Does anyone know why this port is opened, and how to close it? Roshan Punnoose rpunno...@proteuseng.com Proteus Technologies -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Deploying plain spring-dm config file...
Just use spring:xxx where xxx is any valid url (using file, http or mvn) ... On Thu, Nov 11, 2010 at 17:08, Brad Beck bb...@peoplenetonline.com wrote: Is there any support for deploying a plain spring-dm configuration file using osgi:install or as part of a feature xml? Appears the spring deployer currently supports dropping plain configuration files in the deploy directory. I’d like to be able to do this as part of a feature xml. -Brad -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Startup problems: javax.naming.NoInitialContextException: Unable to determine caller'sBundleContext
This exception comes from the aries jndi bundle it seems. I've tested various combinations of karaf and aries jndi, but all seem to have this behavior. COuld you please raise a JIRA issue ? I'm not sure yet, but I suspect the problem is in the aries jndi bundle. On Tue, Nov 30, 2010 at 08:56, Bengt Rodehav be...@rodehav.com wrote: I use Karaf 2.1.2. On a clean startup (the data directory is empty) everything works OK. However, when I restart Karaf (without cleaning out the data directory) I consistently get the following exception: Exception in thread JMX Connector Thread [service:jmx:rmi:///jndi/rmi://localhost:1099/karaf-root] java.lang.RuntimeException: Could not start JMX connector server at org.apache.karaf.management.ConnectorServerFactory$1.run(ConnectorServerFactory.java:103) Caused by: java.io.IOException: Cannot bind to URL [rmi://localhost:1099/karaf-root]: javax.naming.NoInitialContextException: Unable to determine caller's BundleContext at javax.management.remote.rmi.RMIConnectorServer.newIOException(RMIConnectorServer.java:804) at javax.management.remote.rmi.RMIConnectorServer.start(RMIConnectorServer.java:417) at org.apache.karaf.management.ConnectorServerFactory$1.run(ConnectorServerFactory.java:101) Caused by: javax.naming.NoInitialContextException: Unable to determine caller'sBundleContext at org.apache.aries.jndi.OSGiInitialContextFactoryBuilder.getInitialContext(OSGiInitialContextFactoryBuilder.java:53) at javax.naming.spi.NamingManager.getInitialContext(NamingManager.java:667) at javax.naming.InitialContext.getDefaultInitCtx(InitialContext.java:288) at javax.naming.InitialContext.getURLOrDefaultInitCtx(InitialContext.java:316) at javax.naming.InitialContext.bind(InitialContext.java:400) at javax.management.remote.rmi.RMIConnectorServer.bind(RMIConnectorServer.java:625) at javax.management.remote.rmi.RMIConnectorServer.start(RMIConnectorServer.java:412) ... 1 more Does anyone have any idea what could be wrong? I'm thinking there may be timing errors. Note that the exception is only logged to the console - not to the log file. Perhaps it occurs before the logging bundle is installed. I haven't seen this in previous versions of Karaf (2.1.0 and 1.6.0) but I've also added and changed a lot of bundles in addition to upgrading to Karaf 2.1.2. Thus I don't know for sure if this is a problem specific to Karaf 2.1.2. /Bengt -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Ather URL Handler not available.
Yes, those are harmless and can be safely ignored. On Tue, Nov 30, 2010 at 11:58, Bengt Rodehav be...@rodehav.com wrote: Starting from Karaf 2.1.2 I get the following in my log file: 2010-11-30 09:21:45,911 | INFO | rint Extender: 2 | AetherBridgeConnection | .internal.AetherBridgeConnection 61 | Ather URL Handler not available. 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | core | ? ? | ServiceEvent REGISTERED 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintEventDispatcher | ntainer.BlueprintEventDispatcher 123 | Sending blueprint container event BlueprintEvent[type=CREATED] for bundle org.apache.karaf.admin.core 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintContainerImpl | container.BlueprintContainerImpl 231 | Running blueprint container for bundle org.apache.karaf.admin.core in state Created 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintContainerImpl | container.BlueprintContainerImpl 231 | Running blueprint container for bundle org.apache.karaf.deployer.blueprint in state Unknown 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintContainerImpl | container.BlueprintContainerImpl 190 | Grace-period directive: false 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintEventDispatcher | ntainer.BlueprintEventDispatcher 123 | Sending blueprint container event BlueprintEvent[type=CREATING] for bundle org.apache.karaf.deployer.blueprint 2010-11-30 09:21:45,911 | INFO | rint Extender: 2 | AetherBridgeConnection | .internal.AetherBridgeConnection 66 | Using mvn fallback to resolve mvn:se.digia.connect/karaf/1.1.0-SNAPSHOT/xml/features I get lots of these, probably for all the bundles I have installed. Is this normal? What does it mean? /Bengt -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Ather URL Handler not available.
That's in pax-url, see http://issues.ops4j.org/browse/PAXURL-90 2010/11/30 Łukasz Dywicki l...@code-house.org: What do you think to increase log level for this category by default? (Or report decrease proposal in aries) Best regards, Lukasz From: bengt.rode...@gmail.com [mailto:bengt.rode...@gmail.com] On Behalf Of Bengt Rodehav Sent: Tuesday, November 30, 2010 1:38 PM To: user@karaf.apache.org Subject: Re: Ather URL Handler not available. OK - thanks, /Bengt 2010/11/30 Guillaume Nodet gno...@gmail.com Yes, those are harmless and can be safely ignored. On Tue, Nov 30, 2010 at 11:58, Bengt Rodehav be...@rodehav.com wrote: Starting from Karaf 2.1.2 I get the following in my log file: 2010-11-30 09:21:45,911 | INFO | rint Extender: 2 | AetherBridgeConnection | .internal.AetherBridgeConnection 61 | Ather URL Handler not available. 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | core | ? ? | ServiceEvent REGISTERED 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintEventDispatcher | ntainer.BlueprintEventDispatcher 123 | Sending blueprint container event BlueprintEvent[type=CREATED] for bundle org.apache.karaf.admin.core 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintContainerImpl | container.BlueprintContainerImpl 231 | Running blueprint container for bundle org.apache.karaf.admin.core in state Created 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintContainerImpl | container.BlueprintContainerImpl 231 | Running blueprint container for bundle org.apache.karaf.deployer.blueprint in state Unknown 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintContainerImpl | container.BlueprintContainerImpl 190 | Grace-period directive: false 2010-11-30 09:21:45,911 | DEBUG | rint Extender: 3 | BlueprintEventDispatcher | ntainer.BlueprintEventDispatcher 123 | Sending blueprint container event BlueprintEvent[type=CREATING] for bundle org.apache.karaf.deployer.blueprint 2010-11-30 09:21:45,911 | INFO | rint Extender: 2 | AetherBridgeConnection | .internal.AetherBridgeConnection 66 | Using mvn fallback to resolve mvn:se.digia.connect/karaf/1.1.0-SNAPSHOT/xml/features I get lots of these, probably for all the bundles I have installed. Is this normal? What does it mean? /Bengt -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Placing properties files in the classpath
Right, using ConfigAdmin is the way to go in OSGi. If you use blueprint for example, you just have to define a specific proprty placeholder and it's really easy to use. On Thursday, December 9, 2010, Freeman Fang freeman.f...@gmail.com wrote: Hi, You can put properties in $KARAF_HOME/etc folder and reference it from all bundles using ConfigAdmin service. Freeman On 2010-12-9, at 上午7:41, Mike Van wrote: All, The application we are building has several properties files used by all of our bundles. Some of these properties files are .xml, some are plain old .properties files. Currently, I am including them in each bundle, something which makes thier maintenance quite a pain. We would like them available to all bundles, and to have them stored in a single place. So far I have tried placing them in packages which contain code, but the compiler only picks up the .class files. I've also tried to create a package underneath resources named myApp.configurations. This is picked up, but because it isn't exported, it is unavailable for wiring. How does one make a single properties file available to all bundles deployed in Karaf? v/r, Mike Van -- View this message in context: http://karaf.922171.n3.nabble.com/Placing-properties-files-in-the-classpath-tp2054553p2054553.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Freeman Fang FuseSource: http://fusesource.comblog: http://freemanfang.blogspot.comtwitter: http://twitter.com/freemanfangApache Servicemix:http://servicemix.apache.orgApache Cxf: http://cxf.apache.orgApache Karaf: http://karaf.apache.orgApache Felix: http://felix.apache.org -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Placing properties files in the classpath
If you use ConfigAdmin directly or indirectly, you need to specify a configuration id to load the properties from. A single configuration can be used by multiple bundles at the same time, so it's just about using the same id for multiple bundles. If you name the file myApp.cfg, the id of the configuration will be 'myApp'. On Thu, Dec 9, 2010 at 22:52, Mike Van mvangeert...@comcast.net wrote: We were just talking about the differences between bundles we should use as services, and bundles that simply need to be wired. In my definition, all cross-cutting concerns should be services consumed by their bundles. We also have been discussing whether or not the services should all be stateless (I beleive they should be). So, for the time being, if my bundles are all myApp.*, would a myApp.cfg file placed in the etc directory be read by all bundles whose packages start with myApp? - Original Message - From: Łukasz Dywicki [via Karaf] ml-node+2060113-309240584-228...@n3.nabble.com To: Mike Van mvangeert...@comcast.net Sent: Thursday, December 9, 2010 4:06:15 PM Subject: RE: Placing properties files in the classpath No, These bundles may reffer same persistent id (configuration file) without problems. In fact - you may introduce new bundle which produces connection factory and export it as service to reduce number of configuration dependencies. Best regards, Lukasz -Original Message- From: Mike Van [mailto: [hidden email] ] Sent: Thursday, December 09, 2010 10:03 PM To: [hidden email] Subject: Re: Placing properties files in the classpath Ok. If I have 4 bundles that all use JMS, and they are named: myApp.bundle1 myApp.bundle2 myApp.bundle3 myApp.bundle4 Would I need 4 configuration files in etc: myApp.bundle1.cfg myApp.bundle2.cfg myApp.bundle3.cfg myApp.bundle4.cfg ? - Original Message - From: Łukasz Dywicki [via Karaf] [hidden email] To: Mike Van [hidden email] Sent: Thursday, December 9, 2010 3:58:32 PM Subject: RE: Placing properties files in the classpath It depends on the configuration admin. Karaf uses etc directory for these configurations - eg. If you persistence id is set to com.mycompany any changes in $KARAF_BASE/etc/com.mycompany.cfg will be visible for your components. It doesn't look classpath, it looks into etc directory. That's better than classpath because operations can do changes without JAR modification. Even more fantastic is fact that your component can be notified about configuration change.. Best regards, Lukasz -Original Message- From: Mike Van [mailto: [hidden email] ] Sent: Thursday, December 09, 2010 9:42 PM To: [hidden email] Subject: RE: Placing properties files in the classpath In those cases, where does OSGi look to find the properties? And, what are the property file names? Mike Van -- View this message in context: http://karaf.922171.n3.nabble.com/Placing-properties-files-in-the-classpath- tp2054553p2060007.html Sent from the Karaf - User mailing list archive at Nabble.com. View message @ http://karaf.922171.n3.nabble.com/Placing-properties-files-in-the-classpath-tp2054553p2060078.html To start a new topic under Karaf - User, email [hidden email] To unsubscribe from Karaf - User, click here . -- View this message in context: http://karaf.922171.n3.nabble.com/Placing-properties-files-in-the-classpath-tp2054553p2060101.html Sent from the Karaf - User mailing list archive at Nabble.com. View message @ http://karaf.922171.n3.nabble.com/Placing-properties-files-in-the-classpath-tp2054553p2060113.html To start a new topic under Karaf - User, email ml-node+930749-917263437-228...@n3.nabble.com To unsubscribe from Karaf - User, click here . -- View this message in context: http://karaf.922171.n3.nabble.com/Placing-properties-files-in-the-classpath-tp2054553p2060367.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Graceful shutdown of Windows service, revisited
and issue the shutdown command before stopping the service but then it becomes too complicated for most people. Karaf-176 seems to not have solved the problem (for Windows anyway). Shall I reopen that ticket? /Bengt -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: [maven-bundle-plugin] shading version issue
Not sure if that's a typo, but the Export-Package element must be
contained inside the instructions element for the maven plugin to
recognize those.
On Tue, Dec 14, 2010 at 22:16, Mike Van mvangeert...@comcast.net wrote:
When using the maven-bundle-plugin to group non-bundled or improperly bundled
packages into a new bundle, I noticed that the version number of the
resulting packages is set to 0.0.0. Is there any way to fix this? Below
is some code that should allow you to see the problem:
build
plugins
plugin
groupIdorg.apache.felix/groupId
artifactIdmaven-bundle-plugin/artifactId
version2.1.0/version
extensionstrue/extensions
configuration
Export-Package
bad.vendor.package1;version=0.1.2,
bad.vendor.package2;version=3.2.3,
bad.vendor.package3;version=4.3.4
/Export-Package
/configuration
/plugin
/plugins
dependenciesall the bad.vendor dependencies containing the above
packages/dependencies
This results in a file where the packages are not versioned, so when the
resultant bundle is deployed into OSGi, all of the packages are set to
version 0.0.0.
In cases where we are shading together .jar files that are no longer in
development, or .jar files from vendors that simply won't play osgi well, we
create a maven-project and create this bundle along with each versioned
release of our product. As our project grows, contracts, or changes, we
need to add or remove things from this bundle. Because the package versions
are always set to 0.0.0, there is the possibility of ClassNotFound or
constraint violations.
It would be a good idea to have a way to change the behavior of this plugin
to allow the user to set the resultant package versions, or to override the
default behavior to use the ${pom.version} for each package instead of
0.0.0.
v/r,
Mike Van (karafman)
--
View this message in context:
http://karaf.922171.n3.nabble.com/maven-bundle-plugin-shading-version-issue-tp2088176p2088176.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: Web Console Security Question
We do have support for encrypted passwords for JAAS realms, see http://karaf.apache.org/manual/2.1.99-SNAPSHOT/developers-guide/security-framework.html It will be part of 2.2.0 targeted for january. On Wed, Dec 15, 2010 at 17:31, mohamadan mohama...@yahoo.com wrote: Thank you..I just registered for the mailing list. Any estimate when a new Version will support the custom JAAS Login Module ? Also, I noticed that the passwords are in plain text by default, but there is a way to encrypt them. Does it support FIPs-140 encryption ? if not, is it possible to use a different Encryption other than the default one. -- View this message in context: http://karaf.922171.n3.nabble.com/Web-Console-Security-Question-tp2086921p2093005.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: List Range...
Not afaik. One possible way to work around this syntax limitation
would be to define a command range and do something like:
each (range 1 10) { start $it }
Basically, you need a way to compute your set of values to iterate through.
On Mon, Dec 20, 2010 at 15:39, Brad Beck bb...@peoplenetonline.com wrote:
Is there a way to specify a range of values (similar to perl) in the list
operator at the shell? e.g.
each [1..10] { start $it }
-Brad
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: Managed Properties question
Or use a cm:property-placeholder in combination with the reload flag
to reload the app if the config change, that's what i used inside
karaf, see:
http://svn.apache.org/repos/asf/karaf/trunk/shell/ssh/src/main/resources/OSGI-INF/blueprint/shell-ssh.xml
2010/12/20 Łukasz Dywicki l...@code-house.org:
If you try to do managed component your configuration property name must
match field name, otherwise container will not update your bean. Both
spring-dm and aries blueprint works same. In your example the property named
integer have to use placeholder ${integer}. If you would like to use
different names you need to manage changes in bean (set strategy to bean
managed and callback method).
Best regards,
Lukasz
-Original Message-
From: Achim Nierbeck [mailto:bcanh...@googlemail.com]
Sent: Monday, December 20, 2010 6:00 PM
To: user@karaf.apache.org
Subject: Re: Managed Properties question
OK, another try
bean id=containerManaged class=ContainerManagedBean
osgix:managed-properties persistent-id=labX
update-strategy=container-managed/
property name=integer value=23/
/bean
this is the official example :)
I think the property (named integer here) is optional. Usually all
properties which can be read through getter and setter can be set by
the configuration.
So your problem is that you try to inject properties within another property
property name=hibernateProperties
props
prop key=hibernate.show_sql${myAppDbShowSql}/prop
prop key=hibernate.format_sql${myAppDbFormatSql}/prop
/props
/property
All your properties you want to update need to be accessible via
getter and setter.
If you want to do this you need to make an extra bean which is
configurable (with the same pid)
and inject that one after it is initialized. you may want to make your
standard bean dependend on the new bean.
2010/12/20 karafman mvangeert...@comcast.net
It still isn't working.
Here's an excerpt of my .cfg file (none of the names in the file have any
characters other than [a-z, A-Z]:
myAppDbShowSql = false
myAppDbFormatSql = false
In the file where I get my service I have:
osgix:cm-properties id=myAppDatabaseProperties
persistent-id=myApp.data.access/
ctx:property-placeholder properties-ref=myAppDatabaseProperties/
In the file where I am using the managed service I have (unnecessary bits
removed):
bean id=myAppSessionFactory
class=org.springframework.orm.hibernate3.LocalSessionFactoryBean
p:dataSource-ref=myAppPoolDataSource
osgix:managedProperties persistent-id=myApp.data.access
update-strategy=container-managed/
property name=mappingResources
listhibernate valies/list
/property
property name=hibernateProperties
props
prop key=hibernate.show_sql${myAppDbShowSql}/prop
prop key=hibernate.format_sql${myAppDbFormatSql}/prop
/props
/property
/bean
When I change the values in my .cfg file, I can see those changes being
populated by doing a config:list.
However, I get the following error in my log:
Configuration for myApp.data.access has already been used for service
[org.osgi.service.cm.ManagedService, id=127, bundle=89] and will be also
given to [org.osgi.service.cm.ManagedService, id=128, bundle=89]
When I restart bundle 89, the properties are properly consumed.
Anyone know what's going wrong?
-
Karafman
Slayer of the JEE
Pounder of the Perl Programmer
--
View this message in context:
http://karaf.922171.n3.nabble.com/Managed-Properties-question-tp2107407p2121
172.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: features.cfg issue
Just use file:./etc/myApp.features.cfg
The current directory is always ${karaf.base}
But I think it should work anyway as FileInstall do use System
properties to interpolate the values.
Actually, we use it in etc/org.ops4j.pax.url.mvn.cfg, so ther emust be
something wrong with your config.
On Wed, Dec 22, 2010 at 16:41, karafman mvangeert...@comcast.net wrote:
When adding custom features.xml documents to org.apache.karaf.features.cfg on
the featuresRepositories line using the file:/// url, this works:
featuresRepositories=file:///home/myArea/karaf-apache-2.0.0/etc/myApp.features.cfg
But this does not:
featuresRepositories=file://${karaf.base}/etc/myApp.features.cfg
When running config:list, the following is seen:
featuresRepositories
file:///home/myArea/karaf-apache-2.0.0/etc/myApp.features.cfg
The error in the log is:
Caused by: java.net.URISyntaxException: Illegal character in authority at
index 7: file://${karaf.base}/etc/myApp.features.cfg
Karaf should resolve variables in the file url prior to attempting to get a
file-handle, thus allowing users to specify a file url containing
${karaf.base} instead of hard-coding the file location.
-
Karafman
Slayer of the JEE
Pounder of the Perl Programmer
--
View this message in context:
http://karaf.922171.n3.nabble.com/features-cfg-issue-tp2131894p2131894.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: Problem reusing JAAS login modules
It's certainly a missing import package. When drpped into the deploy folder, the deployer will automatically add the required import packages. On Tuesday, December 28, 2010, Rafael Marins rafael.mar...@neociclo.com wrote: Hi, When using the PropertiesLoginModule from a blueprint file in my bundle located at the /OSGI-INF/blueprint/ folder, I've encountered problem with ClassNotFoundException. But I can get it working properly when deploying the blueprint xml into the /deploy/ folder. Check the logging stack trace here: http://pastebin.com/yFiKkpg3 (at line 57) From my blueprint file: accord-jaas-module.xml ... jaas:config name=accord jaas:module className=org.apache.karaf.jaas.modules.properties.PropertiesLoginModule flags=required users = $[karaf.base]/etc/accord-users.properties debug = true /jaas:module /jaas:config ... Any ideas on what must be done to solve this problem? --Rafael Marins -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Branded Karaf...
There's no cleanly defined way to do that. I have in mind since several months to enhance the features maven plugin from Karaf to help doing that in a much easier way, i.e. have a simple way to create custom distributions on top of karaf by overlaying a set of karaf features + configuration files + branding (which could include renaming the main scripts to something different than karaf/karaf.bat) and let the plugin do all the work. There's a jira for that but nobody has had much time to work on that unfortunately. In the mean time, what ServiceMix does is using the maven assembly plugin to add all the required files on top of the unpacked karaf distribution and rebuild archives from that. On Tue, Dec 28, 2010 at 20:41, Kit Plummer kitplum...@gmail.com wrote: Hey Karafers. I'm building a framework on top of Karaf, that is currently only adding the Branding library and a few bundles. But, it is likely in the future I'll need to remove some things, and more. Is there a best way strategy to manage to relationship between my framework and Karaf? How does ServiceMix do it? TIA, Kit -- View this message in context: http://karaf.922171.n3.nabble.com/Branded-Karaf-tp2158363p2158363.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Problems with Karaf hotdeploy
Can you give the output of osgi:headers for the easyb bundle once installed manually ? FileInstall only installs correct jars and r4 compatible bundle. This means that the MANIFEST.MF has to be the first or second entry in the jar (eventually after META-INF/ ) On Thu, Jan 6, 2011 at 13:09, Jürgen Kindler jkind...@talend.com wrote: Hi, I’m experiencing a problem with deploying a bundle into Karaf 2.1.2 by copying it into the karaf_home/deploy folder. Basically I copy three bundles: commons-cli-1.2.jar, groovy-all-1.7.5.jar and easyb-0.9.8.jar I turned the default log level of karaf to debug, but I don’t see any hints about why it fails. In the logs I see that commons-cli-1.2.jar groovy-all-1.7.5.jar are installed and started successfully: 12:34:20,897 | DEBUG | Event Dispatcher | cli | ? ? | 32 - org.apache.commons.cli - 1.2 | BundleEvent INSTALLED 12:34:20,899 | DEBUG | Event Dispatcher | cli | ? ? | 32 - org.apache.commons.cli - 1.2 | BundleEvent RESOLVED 12:34:20,900 | DEBUG | raf-2.1.2/deploy | BlueprintExtender | rint.container.BlueprintExtender 210 | 7 - org.apache.aries.blueprint - 0.2.0.incubating | Scanning bundle org.apache.commons.cli for blueprint application 12:34:20,900 | DEBUG | raf-2.1.2/deploy | BlueprintExtender | rint.container.BlueprintExtender 276 | 7 - org.apache.aries.blueprint - 0.2.0.incubating | No blueprint application found in bundle org.apache.commons.cli 12:34:20,900 | DEBUG | Event Dispatcher | cli | ? ? | 32 - org.apache.commons.cli - 1.2 | BundleEvent STARTED 12:34:31,155 | DEBUG | Event Dispatcher | groovy-all | ? ? | 33 - groovy-all - 1.7.5 | BundleEvent INSTALLED 12:34:31,189 | DEBUG | Event Dispatcher | groovy-all | ? ? | 33 - groovy-all - 1.7.5 | BundleEvent RESOLVED 12:34:31,208 | DEBUG | raf-2.1.2/deploy | BlueprintExtender | rint.container.BlueprintExtender 210 | 7 - org.apache.aries.blueprint - 0.2.0.incubating | Scanning bundle groovy-all for blueprint application 12:34:31,214 | DEBUG | raf-2.1.2/deploy | BlueprintExtender | rint.container.BlueprintExtender 276 | 7 - org.apache.aries.blueprint - 0.2.0.incubating | No blueprint application found in bundle groovy-all 12:34:31,214 | DEBUG | Event Dispatcher | groovy-all | ? ? | 33 - groovy-all - 1.7.5 | BundleEvent STARTED However, there is nothing for the third bundle easyb-0.9.8.jar Note that this really is a bundle and it is possible to explicitly install it using: osgi:install file:/tmp/jki/apache-karaf-2.1.2/deploy/easyb-0.9.8.jar Also starting it with its then assigned bundle id and using it works fine. Somehow it seems like it is ignored by the process that watches the karaf_home/deploy folder :-( Any ideas what the problem is or how I can figure out the root cause. Cheers Juergen -- Jürgen Kindler -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: getting felix scr commands to show up in karaf?
The commands are exposed using plain OSGi, but it should still work without a bueprint descriptor (those are just services in the OSGi registry at the end). I'll have a look at it as there should be a way to get that working. On Mon, Jan 17, 2011 at 08:01, Jean-Baptiste Onofré j...@nanthrax.net wrote: Hi David, The Felix Scr bundle doesn't contain blueprint descriptor (in META-INF/OSGI-INF). Regarding the Import-Package, we can see that it uses org.apache.felix.shell as optional. Without blueprint descriptor, commands won't appear automatically after installation of the bundle. You have to restart your karaf container. Could you make the try ? 1/ install the scr bundle, 2/ restart karaf without deleting the data directory Anyway, we can submit a patch to felix containing scr blueprint. Regards JB On 01/14/2011 07:37 PM, David Jencks wrote: Does anyone know how to get the felix scr commands to show up in the karaf console? I'm trying scr and karaf from trunk and have also installed and started osgi:install -s mvn:org.apache.felix/org.apache.felix.shell/ install file:///Users/david/.m2/repository/org/apache/felix/org.apache.felix.scr/1.6.1-SNAPSHOT/org.apache.felix.scr-1.6.1-SNAPSHOT.jar I expected to see something like scr:list in the list of commands I get fromtab but nothing shows up. Thanks in advance for any help david jencks -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: getting felix scr commands to show up in karaf?
The reason is that the CommandsCompleter (which does completion of command names) only takes into account commands implementing Function (see checkData() method), whereas gogo support reflective commands defined as Methods for example. However, we need to be able to access the mapping between the command object and it's scope / name. On Mon, Jan 17, 2011 at 09:25, Guillaume Nodet gno...@gmail.com wrote: Then it must be a refresh issue somehow. Actually, I've just made some tests and the commands are functional even if they don't appear in the tab completion. That could be considered a Karaf bug. On Mon, Jan 17, 2011 at 08:48, Jean-Baptiste Onofré j...@nanthrax.netwrote: I saw that commands not exposed with Blueprint don't appear automatically after installation. I restarted Karaf to get it working. Regards JB On 01/17/2011 08:41 AM, Guillaume Nodet wrote: The commands are exposed using plain OSGi, but it should still work without a bueprint descriptor (those are just services in the OSGi registry at the end). I'll have a look at it as there should be a way to get that working. On Mon, Jan 17, 2011 at 08:01, Jean-Baptiste Onofré j...@nanthrax.net mailto:j...@nanthrax.net wrote: Hi David, The Felix Scr bundle doesn't contain blueprint descriptor (in META-INF/OSGI-INF). Regarding the Import-Package, we can see that it uses org.apache.felix.shell as optional. Without blueprint descriptor, commands won't appear automatically after installation of the bundle. You have to restart your karaf container. Could you make the try ? 1/ install the scr bundle, 2/ restart karaf without deleting the data directory Anyway, we can submit a patch to felix containing scr blueprint. Regards JB On 01/14/2011 07:37 PM, David Jencks wrote: Does anyone know how to get the felix scr commands to show up in the karaf console? I'm trying scr and karaf from trunk and have also installed and started osgi:install -s mvn:org.apache.felix/org.apache.felix.shell/ install file:///Users/david/.m2/repository/org/apache/felix/org.apache.felix.scr/1.6.1-SNAPSHOT/org.apache.felix.scr-1.6.1-SNAPSHOT.jar I expected to see something like scr:list in the list of commands I get fromtab but nothing shows up. Thanks in advance for any help david jencks -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Console extensions without blueprint
The namespace you use is wrong, it should be http://karaf.apache.org/xmlns/shell/v1.0.0 On Sun, Jan 23, 2011 at 19:18, Adam Crain acr...@greenenergycorp.com wrote: Thanks for the tips. I stripped everything down to the basic example in the manual, and it seems the problem I was having is some kind of blueprint issue. When I run the example console extension, my bundle starts in the GracePeriod and the log diplays: 13:13:04,435 | WARN | rint Extender: 3 | BlueprintContainerImpl | container.BlueprintContainerImpl 252 | 7 - org.apache.aries.blueprint - 0.2.0.incubating | Bundle reef.direct-shell is waiting for namespace handlers [((objectClass=org.apache.aries.blueprint.NamespaceHandler)(osgi.service.blueprint.namespace=http://felix.apache.org/karaf/xmlns/shell/v1.0.0))] ideas? -Adam On Fri, Jan 21, 2011 at 5:01 PM, Guillaume Nodet gno...@gmail.com wrote: Basically, the shell will recognize service that has the osgi.command.scope and osgi.command.function properties defined. However, in order to leverage completions, you currently need to have the service implement AbstractCommand. The easiest way would be to leverage (or adapt) the code in the export() method of the following class: http://svn.apache.org/repos/asf/karaf/trunk/shell/console/src/main/java/org/apache/felix/gogo/commands/basic/SimpleCommand.java On Fri, Jan 21, 2011 at 22:47, Adam Crain acr...@greenenergycorp.com wrote: How do I implement a console extension without blueprint? I created a BundleActivator and tried to publish the service OsgiCommandSupport guessing that would cause the system to pickup the new shell command, but I was wrong. thanks, Adam -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Info about kar archive of Apache Karaf
Yes, kars are for features what eba are for aries applications. Though David Jencks is currently doing a lot of work in this area to have a nice maven integration for kars, so that building a karaf distribution will be done done in a very simply way by referencing kars as maven dependencies and kars will be created as a maven packaging along with features. On Mon, Jan 24, 2011 at 16:46, Charles Moulliard cmoulli...@gmail.com wrote: Is it similar to eba archive of Aries ? On Mon, Jan 24, 2011 at 3:02 PM, Adrian Trenaman adrian.trena...@googlemail.com wrote: The purpose of the Kar file is to facilitate easy packaging and deployment of Karaf features. A feature can have its own bundles *and* all its dependencies placed in a Kar; then, when the Kar archive file is dropped into a deploy directory, the bundles are extracted to the local drive in a psuedo-Maven directory structure, and any features files therein are automatically registered in the runtime. This makes deployment of Karaf solutions easier, particularly when on production machines where Maven resolution is not possible. On 24/01/2011, Charles Moulliard cmoulli...@gmail.com wrote: Hi, What is the purpose of the archive kar that we can deploy now on Karaf ? Benefits ? Regards, Charles Moulliard Sr. Principal Solution Architect - FuseSource Apache Committer Blog : http://cmoulliard.blogspot.com Twitter : http://twitter.com/cmoulliard Linkedin : http://www.linkedin.com/in/charlesmoulliard Skype: cmoulliard -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Apache Karaf WAR + JSF
Also, make sure the jsf-api isn't deployed as an OSGi bundle, else that one could be used and it would not be able to find the implementation inside the jar. On Fri, Feb 11, 2011 at 09:22, Charles Moulliard cmoulli...@gmail.com wrote: The WAR contains the spec and the implementation. aristo-1.0.0.jar barbecue-1.5-beta1.jar bcmail-jdk14-1.38.jar bcmail-jdk14-138.jar bcprov-jdk14-1.38.jar bcprov-jdk14-138.jar bctsp-jdk14-1.38.jar black-tie-1.0.0.jar blitzer-1.0.0.jar bluesky-1.0.0.jar casablanca-1.0.0.jar commons-fileupload-1.2.1.jar commons-io-1.4.jar commons-logging-1.1.1.jar cupertino-1.0.0.jar dark-hive-1.0.0.jar dot-luv-1.0.0.jar eggplant-1.0.0.jar excite-bike-1.0.0.jar facestrace-1.1.0.jar flick-1.0.0.jar hot-sneaks-1.0.0.jar humanity-1.0.0.jar itext-2.1.7.jar jcommon-1.0.0.jar jdom-1.0.jar jfreechart-1.0.0.jar -- jsf-api-2.0.4-b09.jar jsf-impl-2.0.4-b09.jar -- jstl-1.2.jar junit-3.8.jar le-frog-1.0.0.jar log4j-1.2.13.jar midnight-1.0.0.jar mint-choc-1.0.0.jar overcast-1.0.0.jar pepper-grinder-1.0.0.jar poi-3.2-FINAL.jar primefaces-3.0-SNAPSHOT.jar redmond-1.0.0.jar rocket-1.0.0.jar rome-1.0.jar smoothness-1.0.0.jar south-street-1.0.0.jar start-1.0.0.jar sunny-1.0.0.jar swanky-purse-1.0.0.jar trontastic-1.0.0.jar ui-darkness-1.0.0.jar ui-lightness-1.0.0.jar vader-1.0.0.jar On Fri, Feb 11, 2011 at 5:06 AM, David Jencks david_jen...@yahoo.com wrote: You need both the api jar and the implementation in your war. Which jsf implementation? You might consider trying myfaces-bundle deployed outside your web app (removing jsf jars from the web app). This works find in geronimo but I'm not sure at the moment if we do extra initialization to get around this problem. thanks david jencks On Feb 10, 2011, at 5:33 AM, Charles Moulliard wrote: Hi, I have deployed a WAR project on Karaf which is JSF technology based but get this issue : 14:05:19,408 | WARN | FelixStartLevel | war | .eclipse.jetty.util.log.Slf4jLog 50 | 1834 - org.eclipse.jetty.util - 7.2.2.v20101205 | unavailable java.lang.IllegalStateException: Application was not properly initialized at startup, could not find Factory: javax.faces.context.FacesContextFactory at javax.faces.FactoryFinder$FactoryManager.getFactory(FactoryFinder.java:804)[1957:file__Users_charlesmoulliard_Apache_karaf_assembly_target_apache-karaf-2.1.99-SNAPSHOT_deploy_prime-showcase-1.0.0-SNAPSHOT.war:0] at javax.faces.FactoryFinder.getFactory(FactoryFinder.java:306)[1957:file__Users_charlesmoulliard_Apache_karaf_assembly_target_apache-karaf-2.1.99-SNAPSHOT_deploy_prime-showcase-1.0.0-SNAPSHOT.war:0] at javax.faces.webapp.FacesServlet.init(FacesServlet.java:166)[1957:file__Users_charlesmoulliard_Apache_karaf_assembly_target_apache-karaf-2.1.99-SNAPSHOT_deploy_prime-showcase-1.0.0-SNAPSHOT.war:0] at org.eclipse.jetty.servlet.ServletHolder.initServlet(ServletHolder.java:432)[1842:org.eclipse.jetty.servlet:7.2.2.v20101205] at org.eclipse.jetty.servlet.ServletHolder.doStart(ServletHolder.java:260)[1842:org.eclipse.jetty.servlet:7.2.2.v20101205] at org.eclipse.jetty.util.component.AbstractLifeCycle.start(AbstractLifeCycle.java:55)[1834:org.eclipse.jetty.util:7.2.2.v20101205] The lib directory of the WAR contains well the jar of jsf api -- jsf-api-2.0.4-b09 Remark : the same WAR deployed on Tomcat works fine Is it something that you already experienced ? Regards, Charles Moulliard Sr. Principal Solution Architect - FuseSource Apache Committer Blog : http://cmoulliard.blogspot.com Twitter : http://twitter.com/cmoulliard Linkedin : http://www.linkedin.com/in/charlesmoulliard Skype: cmoulliard -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Problems with features startlevel
discussions on this topic? /Bengt 2011/2/12 Guillaume Nodet gno...@gmail.com On Sat, Feb 12, 2011 at 00:38, Bengt Rodehav be...@rodehav.com wrote: Ok, thanks. Is there know way to avoid having to modify startup.properties then? Not really at startup. I think after a restart, things are different as the cglib bundle is already installed and should be used by the framework. (Do you never sleep or are we in different time zones?) GMT +1, and not much / enough sleep unfortunately ;-) Den 12 feb 2011 00.15 skrev Guillaume Nodet gno...@gmail.com: -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Passing negative integer as first command parameter...
Good question. Not really sure actually. Can you try range ' -1' 10 with a space between the first quote and the minus sign ? On Thu, Feb 17, 2011 at 20:15, Brad Beck bb...@peoplenetonline.com wrote: How do I escape the first parameter to a command when it is a negative number (e.g. range -1 10)? I'm currently getting the following error when I try the example above (Karaf 2.1.0)... Error executing command pfm:range undefined option -1 -Brad Beck -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Race between Features and Spring/Blueprint XML
I've seen lots of problems when using spring-dm, some of them do not have any clean solution *at all*, especially when using custom namespace handlers such as camel. That's why I've been advocating for using blueprint instead, which works way better. Camel support for blueprint is much better since 2.6 and 2.7 will bring another set of improvements, so if that's an option, i'd switch to blueprint. Else, well, you can always try to put some Thread.sleep at some very fined tuned location in order to make sure the camel namespace handler is ready and that all camel components are available. On Wed, Mar 9, 2011 at 15:51, Michael Prieß mailingliste...@googlemail.com wrote: Hi all, i like to deploy camel-routes and features inside the deploy directory. A deployment look like the following: - a feature.xml which contain bundles like Apache Camel, Spring with a start level definition. - and many xml camel-routes which contain configurations for my components. Now i have the problem that my camel-routes have the same start level like the features. Have anyone a good idea how to resolve the problem? Regards, Michael Priess -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Help me! about Karaf with spring-DM webApp
Could you reformat your code ? I think a lot has been lost. Also, this is more a question for the users list, so I'm answering there. Last, but not least, which version of Karaf do you use ? If you don't use 2.2.0 which has been released, maybe you could try that one ? On Thu, Mar 10, 2011 at 09:55, stream stream1...@gmail.com wrote: there is a problem that a java.lang.NullPointerException always happen when i used spring-dm in my webApp project. unless deleted the code contextClass org.springframework.osgi.web.context.support.OsgiBundleXmlWebApplicationContext org.springframework.web.context.ContextLoaderListener in my web.xml.. who can help me i have spent three day... exception Could not start the servlet context for http context [org.ops4j.pax.web.extender.war.internal.WebAppWebContainerContext@cb5d35] java.lang.NullPointerException at org.ops4j.pax.web.service.jetty.internal.JettyServerWrapper.addContext(JettyServerWrapper.java:209)[86:org.ops4j.pax.web.pax-web-jetty:1.0.1] at org.ops4j.pax.web.service.jetty.internal.JettyServerWrapper.getOrCreateContext(JettyServerWrapper.java:112)[86:org.ops4j.pax.web.pax-web-jetty:1.0.1] at org.ops4j.pax.web.service.jetty.internal.JettyServerImpl.addServlet(JettyServerImpl.java:137)[86:org.ops4j.pax.web.pax-web-jetty:1.0.1] at org.ops4j.pax.web.service.jetty.internal.ServerControllerImpl$Started.addServlet(ServerControllerImpl.java:266)[86:org.ops4j.pax.web.pax-web-jetty:1.0.1] at org.ops4j.pax.web.service.jetty.internal.ServerControllerImpl.addServlet(ServerControllerImpl.java:107)[86:org.ops4j.pax.web.pax-web-jetty:1.0.1] at org.ops4j.pax.web.service.internal.HttpServiceStarted.registerResources(HttpServiceStarted.java:180)[85:org.ops4j.pax.web.pax-web-runtime:1.0.1] at org.ops4j.pax.web.service.internal.HttpServiceProxy.registerResources(HttpServiceProxy.java:66)[85:org.ops4j.pax.web.pax-web-runtime:1.0.1] at org.ops4j.pax.web.extender.war.internal.RegisterWebAppVisitorWC.visit(RegisterWebAppVisitorWC.java:138)[89:org.ops4j.pax.web.pax-web-extender-war:1.0.1] at org.ops4j.pax.web.extender.war.internal.model.WebApp.accept(WebApp.java:558)[89:org.ops4j.pax.web.pax-web-extender-war:1.0.1] at org.ops4j.pax.web.extender.war.internal.WebAppPublisher$HttpServiceListener.register(WebAppPublisher.java:170)[89:org.ops4j.pax.web.pax-web-extender-war:1.0.1] -- View this message in context: http://karaf.922171.n3.nabble.com/Help-me-about-Karaf-with-spring-DM-webApp-tp2659667p2659667.html Sent from the Karaf - Dev mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: karaf shell extension with iPojo
You can't invoke the command ? or the command does not appear in the
completion when you press tab ?
The former sounds like a bug, but the later is kinda expected.
On Thu, Mar 17, 2011 at 13:48, Lindley Andrew andrew.lind...@ait.ac.at wrote:
Dear all,
I am using Karaf 2.1.0 and want to extend the shell with my own
shell-commands.
There’s a great tutorial showing how to do this with blueprint.
http://karaf.apache.org/manual/2.1.99-SNAPSHOT/developers-guide/extending-console.html
I was trying to do this with iPojo but did not succeed. That’s the sample I
was trying to run.
public interface SampleTUI {
// the supported operations on the shell we're extending
String FUNCTION_STR = [list];
// will be used in the activator to define the namespace within the
shell
String SCOPE = preserv;
public void list();
}
@Component(name=sample.addons.api.command.KarafSampleTUI)
@Provides
public class KarafSampleTUI implements SampleTUI {
// the supported operations on the shell we're extending
@ServiceProperty(name=osgi.command.function,
value=ServiceRegistryTUI.FUNCTION_STR)
public String[] functions;
// will be used in the activator to define the namespace within the
shell
@ServiceProperty(name=osgi.command.scope,
value=ServiceRegistryTUI.SCOPE)
public String scope;
//these fields are injected
//@Requires
//private MyUtils mu;
@Descriptor(some sample description)
public void list() {
System.out.println(testing list);
}
}
ipojo xmlns:xsi=http://www.w3.org/2001/XMLSchema-instance;
xsi:schemaLocation=org.apache.felix.ipojo
http://felix.apache.org/ipojo/schemas/CURRENT/core.xsd
org.apache.felix.ipojo.extender
http://felix.apache.org/ipojo/schemas/CURRENT/extender-pattern.xsd;
xmlns=org.apache.felix.ipojo
instance component=sample.addons.api.command.KarafSampleTUI/
/ipojo
It perfectly runs within Felix using gogo, but not in karaf (using Felix).
Is this specific to gogo?
I also had a look at apache\felix\ipojo\arch\ and
apache\felix\ipojo\arch\gogo but both aren’t working in karaf as well.
Thanks for your support,
Kr Andrew
--
Cheers,
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: How does features:refreshUrl handles snapshot
I'll experiment to see if it is possible to find if the snapshot has changed. Could you please raise a JIRA issue for that ? On Thu, Mar 17, 2011 at 17:27, Guillaume Nodet gno...@gmail.com wrote: First refreshUrl only reload the features descriptors. It doesn't update the features. We've added the dev:watch command in karaf 2.2.0 which can automatically update bundles if newer snapshots are available in your local repo. I guess the problem is to detect when snapshots have actually changed, but I suppose it could be done by looking at the maven metadata. Though it might be a bit more costly than checking the local file system as this would need several http requests for each snapshot, so a sufficient delay should be used between polls. On Thu, Mar 17, 2011 at 17:12, Dan Tran dant...@gmail.com wrote: Hello, I am able to install my snaphot features from my company maven repository. When there is new snapshot on my repo, features:refreshUrl is able to identify the latest snapshot ..-features.xml ( via latest timestamp file ), however after that nothing happens. I am expecting all my SNAPSHOT bundles belong to my features.xml to get downloaded and restart. is this a bug? -Dan -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Followup on getting felix scr commands to show up in karaf
As you noticed, scr 1.6.0 use the old felix shell which is different from gogo, so that won't work unless there's a bridge somewhere. For scr trunk, not sure why it doesn't work. Can you first check if the command is actually registered in the registry using the osgi:ls karaf command ? If the command is registered but isn't available, that's a bug in Karaf console. If the command isn't registered, a suspect an exception to be thrown in the ScrCommand#register() method. On Sat, Mar 19, 2011 at 14:56, Christoper Blunck ch...@thebluncks.com wrote: Hmmm... I don't believe I'm trying to use a karaf feature on top of a gogo command. I'm just trying to run the scr:list command without using tab completion or anything. Steps to reproduce: 1.) Download and untar: http://www.apache.org/dyn/closer.cgi/karaf/2.2.0/apache-karaf-2.2.0.tar.gz 2.) Start up karaf bin running cd bin; ./karaf 3.) Download into the deploy directory: http://www.gossipcheck.com/mirrors/apache//felix/org.apache.felix.scr-1.6.0.jar 4.) type scr:list or scr list at the prompt. You'll get command not found. A couple of things I've noticed... Here's the source code for the bundle activator for SCR in version 1.6.0: http://svn.apache.org/repos/asf/felix/releases/org.apache.felix.scr-1.6.0/src/main/java/org/apache/felix/scr/impl/Activator.java Note that in the start() method there's a try-catch around the part where the ScrCommand is registered. I added a debug statement to the catch section and found out that there was a NoClassDefFoundError on org.apache.felix.shell.Command when the code was trying to register the ScrCommand. I downloaded org.apache.felix.shll.command 1.4.2 and put it my deploy directory. That made the NoClassDefFoundError go away but it still did not make the scr list command work. I still got a command not found. Also ... It appears the trunk of SCR has been updated to work with gogo: http://svn.apache.org/repos/asf/felix/trunk/scr/src/main/java/org/apache/felix/scr/impl/Activator.java The start() method looks different and the comment suggests it now works with gogo. I pulled the trunk and compiled it. I then tried to run karaf with org.apache.felix.scr-1.6.1-SNAPSHOT. DS started but I still didn't have my scr list command. Any ideas? -c On Sat, Mar 19, 2011 at 3:59 AM, Guillaume Nodet gno...@gmail.com wrote: Let's be clear about that. There's no reason why the scr or any gogo commands would not work in karaf. What does not work is karaf features on top of gogo such as completion (both command and parameter). If that's not the case, this should clearly be fixed asap. On Sat, Mar 19, 2011 at 02:33, Christoper Blunck ch...@thebluncks.com wrote: Hello all, Over in this thread the gogo'ness of SCR is discussed: http://karaf.922171.n3.nabble.com/getting-felix-scr-commands-to-show-up-in-karaf-td2257486.html I'm stuck in the same problem where my SCR bundle loads and all my services are properly satisfied and injected but the scr command itself is not available. Guillame you remarked: Actually, I've just made some tests and the commands are functional even if they don't appear in the tab completion I was hoping you could elaborate a little more about this. Are you saying that you were able to get to the scr command in the karaf prompt? Or were you simply remarking that scr worked wrt injection and that the commands just weren't functional? I recognize that this ticket is still out there: https://issues.apache.org/jira/browse/KARAF-397 And I see the priority is MAJOR. Question to the devs: is this something you expect will be fixed soon? I want to go to Karaf 2.2.0 but the lack of a scr command is going to give my developers a lot of heartache Thanks for your time, -c -- Christopher Blunck ch...@thebluncks.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Christopher Blunck ch...@thebluncks.com -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Followup on getting felix scr commands to show up in karaf
That's the expected behavior. Command themselves work but not completion. On Sun, Mar 20, 2011 at 02:35, Christoper Blunck ch...@thebluncks.com wrote: Hi Guillaume- Thanks for giving me some troubleshooting techniques. I ran the osgi:ls command and here is a snippit of the output: Apache Felix Declarative Services (41) provides: org.osgi.service.cm.ConfigurationListener org.apache.felix.scr.ScrService org.osgi.service.cm.ManagedService org.apache.felix.scr.impl.ScrGogoCommand That last line suggests to me that the command itself has been registered. Interestingly enough scr list fails but scr:list succeeds: karaf@root scr list Command not found: scr karaf@root scr:list No components registered When I do a tab-tab at the karaf@root prompt I see a bunch of commands I can execute (which is expected). However, scr:list is not in that list. Here are the commands starting with s that are available to me: set shell:cat shell:clear shell:each shell:echo shell:exec shell:grep shell:head shell:history shell:if shell:info shell:java shell:logout shell:more shell:new shell:printf shell:sleep shell:sort shell:tac shell:tail show-tree shutdown sleep sort ssh ssh:ssh ssh:sshd sshd start start-level stop I am really in over my head wrt gogo and the rest of the shell framework stuff so I'm not quite sure what to do next. Do you have some ideas of what is going on? I'd be happy to help you troubleshoot some more but I don't know how much more valuable I'd be at this point... Thanks again for your time - I do appreciate it. -c On Sat, Mar 19, 2011 at 3:50 PM, Guillaume Nodet gno...@gmail.com wrote: As you noticed, scr 1.6.0 use the old felix shell which is different from gogo, so that won't work unless there's a bridge somewhere. For scr trunk, not sure why it doesn't work. Can you first check if the command is actually registered in the registry using the osgi:ls karaf command ? If the command is registered but isn't available, that's a bug in Karaf console. If the command isn't registered, a suspect an exception to be thrown in the ScrCommand#register() method. On Sat, Mar 19, 2011 at 14:56, Christoper Blunck ch...@thebluncks.com wrote: Hmmm... I don't believe I'm trying to use a karaf feature on top of a gogo command. I'm just trying to run the scr:list command without using tab completion or anything. Steps to reproduce: 1.) Download and untar: http://www.apache.org/dyn/closer.cgi/karaf/2.2.0/apache-karaf-2.2.0.tar.gz 2.) Start up karaf bin running cd bin; ./karaf 3.) Download into the deploy directory: http://www.gossipcheck.com/mirrors/apache//felix/org.apache.felix.scr-1.6.0.jar 4.) type scr:list or scr list at the prompt. You'll get command not found. A couple of things I've noticed... Here's the source code for the bundle activator for SCR in version 1.6.0: http://svn.apache.org/repos/asf/felix/releases/org.apache.felix.scr-1.6.0/src/main/java/org/apache/felix/scr/impl/Activator.java Note that in the start() method there's a try-catch around the part where the ScrCommand is registered. I added a debug statement to the catch section and found out that there was a NoClassDefFoundError on org.apache.felix.shell.Command when the code was trying to register the ScrCommand. I downloaded org.apache.felix.shll.command 1.4.2 and put it my deploy directory. That made the NoClassDefFoundError go away but it still did not make the scr list command work. I still got a command not found. Also ... It appears the trunk of SCR has been updated to work with gogo: http://svn.apache.org/repos/asf/felix/trunk/scr/src/main/java/org/apache/felix/scr/impl/Activator.java The start() method looks different and the comment suggests it now works with gogo. I pulled the trunk and compiled it. I then tried to run karaf with org.apache.felix.scr-1.6.1-SNAPSHOT. DS started but I still didn't have my scr list command. Any ideas? -c On Sat, Mar 19, 2011 at 3:59 AM, Guillaume Nodet gno...@gmail.com wrote: Let's be clear about that. There's no reason why the scr or any gogo commands would not work in karaf. What does not work is karaf features on top of gogo such as completion (both command and parameter). If that's not the case, this should clearly be fixed asap. On Sat, Mar 19, 2011 at 02:33, Christoper Blunck ch...@thebluncks.com wrote: Hello all, Over
Re: Karaf 2.2.0 Startup with no internet Access
That's a bug imho, that file should either not be referenced or be included. Could you please raise a JIRA issue for that ? On Wed, Mar 30, 2011 at 16:43, Hervé BARRAULT herve.barra...@gmail.com wrote: HI, i have download the karaf 2.2.0 and i try to start it. (In my case karaf is installed in a clean machine without any maven installation nor internel access). It is impossible to start it because it seems that the features file is not in the delivery. 16:04:43,527 | WARN | rint Extender: 1 | FeaturesServiceImpl | res.internal.FeaturesServiceImpl 911 | 23 - org.apache.karaf.features.core - 2.2.0 | Unable to add features repository mvn:org.apache.karaf.assemblies.features/enterprise/2.2.0/xml/features at startup java.lang.RuntimeException: URL [mvn:org.apache.karaf.assemblies.features/enterprise/2.2.0/xml/features] could not be resolved. at org.ops4j.pax.url.mvn.internal.Connection.getInputStream(Connection.java:195) at org.apache.karaf.features.internal.FeatureValidationUtil.validate(FeatureValidationUtil.java:49) at org.apache.karaf.features.internal.FeaturesServiceImpl.validateRepository(FeaturesServiceImpl.java:199) at org.apache.karaf.features.internal.FeaturesServiceImpl.internalAddRepository(FeaturesServiceImpl.java:210) at org.apache.karaf.features.internal.FeaturesServiceImpl.start(FeaturesServiceImpl.java:909) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)[:1.6.0_16] at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)[:1.6.0_16] at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)[:1.6.0_16] at java.lang.reflect.Method.invoke(Method.java:597)[:1.6.0_16] at org.apache.aries.blueprint.utils.ReflectionUtils.invoke(ReflectionUtils.java:226)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BeanRecipe.invoke(BeanRecipe.java:824)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BeanRecipe.runBeanProcInit(BeanRecipe.java:636)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BeanRecipe.internalCreate(BeanRecipe.java:724)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.di.AbstractRecipe.create(AbstractRecipe.java:64)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BlueprintRepository.createInstances(BlueprintRepository.java:219)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BlueprintRepository.createAll(BlueprintRepository.java:147)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BlueprintContainerImpl.instantiateEagerComponents(BlueprintContainerImpl.java:640)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BlueprintContainerImpl.doRun(BlueprintContainerImpl.java:331)[10:org.apache.aries.blueprint:0.3.0] at org.apache.aries.blueprint.container.BlueprintContainerImpl.run(BlueprintContainerImpl.java:227)[10:org.apache.aries.blueprint:0.3.0] at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:441)[:1.6.0_16] at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303)[:1.6.0_16] at java.util.concurrent.FutureTask.run(FutureTask.java:138)[:1.6.0_16] at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$301(ScheduledThreadPoolExecutor.java:98)[:1.6.0_16] at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:207)[:1.6.0_16] at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)[:1.6.0_16] at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)[:1.6.0_16] at java.lang.Thread.run(Thread.java:619)[:1.6.0_16] I looked to the system directory and in org\apache\karaf, i have only the following directories : admin, deployer, diagnostic, features, jaas, org.apache.karaf.management, shell (no assemblies). I have also search for xml files and i didn't find one. (I will do some tests with this one : http://repo2.maven.org/maven2/org/apache/karaf/assemblies/features/enterprise/2.2.0/enterprise-2.2.0-features.xml) Am i wrong ? Thanks for answers Herve -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Using custom log4j appenders under Karaf 2.1.4
On Tue, Apr 5, 2011 at 23:43, mgardiner gardin...@familysearch.org wrote: Hi, Are there any examples available showing how to use custom log4j appenders under Karaf? I see in the users guide the following note: If you plan to use your own appenders, you need to create an OSGi bundle and attach it as a fragment to the bundle with a symbolic name of org.ops4j.pax.logging.pax-logging-service. This way, the underlying logging system will be able to see and use your appenders. We have a custom remoting logging appender we wish to utilize with our project hosted in Karaf 2.1.4. I am assuming we turn our appenders jar into a bundle with the fragment host set to org.ops4j.pax.logging.pax-logging-service and deployed with our project. Is that correct? Yep, that's correct. How do you recommend handling multiple log4j.xml files for each deployment environment such as development, staging, and production. Good question. We don't usually use xml files for log4j but rather the configuration based ones. Take a look at the etc/org.ops4j.pax.logging.cfg file which is actually a log4j config file. That file will be used to configure log4j, so you should only touch that file and configure it differently between your environments I think; Thanks. -Mike- -- View this message in context: http://karaf.922171.n3.nabble.com/Using-custom-log4j-appenders-under-Karaf-2-1-4-tp2781811p2781811.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: WAITING on?
log:debug | grep BlueprintEvent It usually helps, but you may need to be at debug level. We definitely need to provide a better way to find this information. Feel free to raise a JIRA for that. On Wed, Apr 13, 2011 at 20:56, mikevan mvangeert...@comcast.net wrote: Is there a console command that will allow you to see what a bundle that is active and Waiting is waiting on? - Mike Van (aka karafman) Karaf Team (Contributor) -- View this message in context: http://karaf.922171.n3.nabble.com/WAITING-on-tp2817312p2817312.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Karaf shell problem with long lines
Yes, and that i've fixed those issues trunk and 2.2.x branch, though that's not released yet. On Wed, Apr 20, 2011 at 23:56, Cristiano Gavião cvgav...@gmail.com wrote: Hi, I'm experiencing some weird behavior using karaf shell command with terminal's macosx. When I tried to install a feature with a long line pasted inside one terminal window (with the size smaller than the copied line) I got a RuntimeException... See: I copied this from a text editor: features:addurl mvn:org.jbehave.osgi/jbehave-osgi-karaf-features/1.0-SNAPSHOT/xml/features and paste it on Karaf command shell... the text is truncated karaf@root features:addurl mvn:org.jbehave.osgi/jbehave-osgi-karaf-features/1.0-SNAPSHOT/xml/featur Could not add Feature Repository: java.lang.RuntimeException: URL [mvn:org.jbehave.osgi/jbehave-osgi-karaf-features/1.0-SNAPSHOT/xml/featur] could not be resolved. But if I enlarge the window size and try again, everything works ok... :-D Could this be a Terminal's or Karaf's problem ? cheers Cristiano -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com Connect at CamelOne May 24-26 The Open Source Integration Conference http://camelone.com/
Re: Installing custom security provider
If boot delegation is available for a given package, it will always be used first. On Wed, Apr 27, 2011 at 10:21, Zhemzhitsky Sergey sergey_zhemzhit...@troika.ru wrote: Hi there, I’m trying to install custom security provider using this http://felix.apache.org/site/65-deploying-security-providers.html link. Unfortunately this provider uses xmlsec library which must be placed near it. I have changed org.osgi.framework.bootdelegation like this: org.osgi.framework.bootdelegation=org.apache.karaf.jaas.boot,sun.*,com.sun.*,javax.transaction,javax.transaction.*, \ org.customsec.*, \ org.apache.xml.security.* And now I need all the bundles to delegate requests to org.apache.xml.security from the org.osgi.framework.bootdelegation even if they import packages from org.apache.xml.security with specific versions. Is it possible to achieve? Best Regards, Sergey Zhemzhitsky ___ The information contained in this message may be privileged and conf idential and protected from disclosure. If you are not the original intended recipient, you are hereby notified that any review, retransmission, dissemination, or other use of, or taking of any action in reliance upon, this information is prohibited. If you have received this communication in error, please notify the sender immediately by replying to this message and delete it from your computer. Thank you for your cooperation. Troika Dialog, Russia. If you need assistance please contact our Contact Center (+7495) 258 0500 or go to www.troika.ru/eng/Contacts/system.wbp -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com Connect at CamelOne May 24-26 The Open Source Integration Conference http://camelone.com/
Re: Continuous Delivery
One thing I've been working on recently is a deployment agent for fabric (see http://gnodet.blogspot.com/2011/05/introducing-fusesource-fabric.html). We've used a slightly different mechanism as the agent is responsible for all the deployments so that it can actually upgrade all the bundles, including karaf bundles and even the osgi framework itself. This way, the agent can actually upgrade features to newer versions. The downside is that you need to go through the agent for all deployments. On Wed, May 4, 2011 at 17:42, bbolingbroke bolingbrok...@familysearch.org wrote: If your goal is to have zero downtime, then when you upgrade to 1.1, would you follow these steps? You release 1.1 and deploy it to the maven repo features:addurl mvn:org.yourorg/yourproject-features/1.1/xml features:install yourapp/1.1 features:uninstall yourapp/1.0 features:removeurl mvn:org.yourorg/yourproject-features/1.0/xml -- View this message in context: http://karaf.922171.n3.nabble.com/Continuous-Delivery-tp2899218p2899341.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com Connect at CamelOne May 24-26 The Open Source Integration Conference http://camelone.com/
Re: Show karaf prompt after bundle println
The prompt is displayed when the execution of a command is finished. If you're still executing the command, the prompt won't be dipslayed, but you can retrieve it from the shell's session variables and display it your self if you need. On Sat, May 14, 2011 at 14:32, Samuel Cox crankydi...@gmail.com wrote: I tried that and it didn't work. I will say that my bean doing the output is written in Scala. I'm using println(), which I think is equivalent to Java's System.out.println(). On Sat, May 14, 2011 at 12:38 AM, Jean-Baptiste Onofré j...@nanthrax.net wrote: Hi Samuel, It should be the case. Karaf shell intercept the std and err output stream. It means that after a System.out.println(), the Karaf prompt should be display. It's the case in the commands. For instance, osgi:list iterates on the bundles and simply display the bundles attribute using System.out.println(). Try to add a System.out.println() at the end of your bundle output, it should display the prompt just after your bundle output. Regards JB On 05/14/2011 03:10 AM, Samuel Cox wrote: Hi, Is it possible to get the prompt back after some bundle code does a System.out.println()? I have some beans that need to print stuff to the Karaf console. I can't figure out how to get the prompt back without the user hitting enter. Many thanks. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com Connect at CamelOne May 24-26 The Open Source Integration Conference http://camelone.com/
Re: Show karaf prompt after bundle println
As I explained the prompt only deals with commands. If you print anything on the console and if you want the prompt to be displayed again. I think the only way would be to create a shell session, and then print the prompt that can be retrieved by calling Console#getPrompt(), see http://svn.apache.org/viewvc/karaf/trunk/shell/console/src/main/java/org/apache/karaf/shell/console/jline/Console.java?revision=1096717view=markup I don't really see any other way for now. Maybe another way would be for you to hack the etc/shell.init.script and perform you diagnostics there ? The script is executed between the welcome lines and the first prompt... On Tue, May 17, 2011 at 15:13, Samuel Cox crankydi...@gmail.com wrote: Shameless bump to give this one more chance;) I've been reading the documentation on custom commands, but nothing is jumping out... Just to be clear on what I need. I need to automatically run diagnostics after all my OSGi bundles have loaded. The results need to be displayed to the user in Karaf's console. The running of the diagnostics is currently be triggered by Spring bean loading (using an init-method). In the past, it was in a bundle activator. In neither case, does the karaf prompt show up unless the user hits the Enter key. It's not the end of the world if I can't get this to work, but it would be nice. On Sat, May 14, 2011 at 8:10 AM, Samuel Cox crankydi...@gmail.com wrote: I'm not executing a command. We do have our own commands, and they work fine. This is just a bean in some bundle that displays some diagnostics on startup. If there is a way to execute a command automatically when a bundle starts, I could print the diagnostics using that. I couldn't find any info on doing that. Thanks for the help btw! On Sat, May 14, 2011 at 7:36 AM, Guillaume Nodet gno...@gmail.com wrote: The prompt is displayed when the execution of a command is finished. If you're still executing the command, the prompt won't be dipslayed, but you can retrieve it from the shell's session variables and display it your self if you need. On Sat, May 14, 2011 at 14:32, Samuel Cox crankydi...@gmail.com wrote: I tried that and it didn't work. I will say that my bean doing the output is written in Scala. I'm using println(), which I think is equivalent to Java's System.out.println(). On Sat, May 14, 2011 at 12:38 AM, Jean-Baptiste Onofré j...@nanthrax.net wrote: Hi Samuel, It should be the case. Karaf shell intercept the std and err output stream. It means that after a System.out.println(), the Karaf prompt should be display. It's the case in the commands. For instance, osgi:list iterates on the bundles and simply display the bundles attribute using System.out.println(). Try to add a System.out.println() at the end of your bundle output, it should display the prompt just after your bundle output. Regards JB On 05/14/2011 03:10 AM, Samuel Cox wrote: Hi, Is it possible to get the prompt back after some bundle code does a System.out.println()? I have some beans that need to print stuff to the Karaf console. I can't figure out how to get the prompt back without the user hitting enter. Many thanks. -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com Connect at CamelOne May 24-26 The Open Source Integration Conference http://camelone.com/ -- Cheers, Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com Connect at CamelOne May 24-26 The Open Source Integration Conference http://camelone.com/
Re: Bundle dependencies
-features-to-repo/goal
/goals
configuration
descriptors
descriptormvn:org.apache.karaf.assemblies.features/standard/2.2.1/xm
l/features/descriptor
descriptormvn:com.myapps/features/${myapps.release}/xml/features/de
scriptor
/descriptors
features
featureconfig/feature
featuressh/feature
featuremanagement/feature
featurejetty/feature
featurewebconsole/feature
featurewar/feature
featurespring/feature
featurespring-web/feature
featurespring-dm/feature
featurespring-dm-web/feature
featuremy-custom-bundle/feature
/features
includeMvnBasedDescriptorstrue/includeMvnBasedDescriptors
repositorytarget/features-repo/repository
addTransitiveFeaturestrue/addTransitiveFeatures
/configuration
/execution
/executions
/plugin
I tried to add the includeMvnBasedDescriptors and
addTransitiveFeatures tags, but it doesn't seem to fix the problem.
I also customized the org.ops4j.pax.ulr.mvn.cfg file, so that the
proxy support is enabled and that the 'org.ops4j.pax.url.mvn.settings'
and 'org.ops4j.pax.url.mvn.localRepository' properties are correctly
set to my local repository and maven settings (where my proxy is defined).
My feature description looks like:
feature name=myFeature version=${myapps.release}
resolver=(orb)
bundlemvn:org.springframework.ldap/org.springframework.ldap/1.3.0.RE
LEASE
/bundle
bundlemvn:com.thoughtworks.xstream/com.springsource.com.thoughtworks
.xstream/1.3.1/bundle
bundlemvn:com.myapps/myapps/${myapps.release}/bundle
/feature
The EFG Mail Gateway made the following annotation This
e-mail is confidential. If you are not the intended recipient, you
should not copy it, re-transmit it, use it or disclose its contents,
but should return it to the sender immediately and delete the copy
from your system.
EFG is not responsible for, nor endorses, any opinion, recommendation,
conclusion, solicitation, offer or agreement or any information
contained in this communication.
EFG cannot accept any responsibility for the accuracy or completeness
of this message as it has been transmitted over a public network.
If you suspect that the message may have been intercepted or amended,
please call the sender. Should you require any further information,
please contact the Compliance Manager on off...@efggroup.com.
==
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
Open Source SOA
http://fusesource.com
Re: Running Karaf without PAX Logging
Most of Karaf only depends on the slf4j api. The only dependencies on pax-logging are related to the management of the logging system itself (which is understandably tied to the logging system implementation). if you remove the shell log commands, you should have no problems. On Sun, Jul 17, 2011 at 22:40, codeoncoffee codeoncof...@gmail.com wrote: Hello, I'm trying to run Karaf hosted within a larger application and just can't work my way around the class-loading issues with the embedded Log4J classes in the PAX Logging implementation. Let me take that back, I can work around them, but this whole OSGI layer is being implemented as a plugin to a larger application. So I have no control to let classes be added to the bootstrap classloader and import them into the OSGI instance. Earlier this year I tried removing PAX and using the simple Felix Log Service. I found several areas where Karaf was dependent on PAX logging being in place. So my question I guess is, Is there any interest in having Karaf be agnostic to Log Service implementation? Any idea how large of a task this would be? I don't mind contributing back, but at this point I'm very close to moving back to raw Felix. -- View this message in context: http://karaf.922171.n3.nabble.com/Running-Karaf-without-PAX-Logging-tp3177504p3177504.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Running Karaf without PAX Logging
The only real use case for not using pax-logging I've heard about is because people wanted to use logback as a backend. I think this might be solved by having a pax-logging implementation based on logback instead. On Mon, Jul 25, 2011 at 04:01, Andreas Pieber anpie...@gmail.com wrote: Just for interest; any reports by now on this topic? This might be interesting for other users too and I think a page in the usermanual (How to use Karaf without pax-logging) might be a good idea. Kind regards, Andreas On Sun, Jul 17, 2011 at 23:01, codeoncoffee codeoncof...@gmail.com wrote: Excellent. I'll give it a try and report back with steps taken -- View this message in context: http://karaf.922171.n3.nabble.com/Running-Karaf-without-PAX-Logging-tp3177504p3177529.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Coming to Grips with Blueprint
As said, those elements are custom elements made available by projects to extend the standard blueprint element set with more specific ones. Each project (camel, activemq, cxf, ...) should document its own namespaces. Karaf provides a few also (jaas, shell). On Sat, Sep 10, 2011 at 00:47, Geoffry Roberts geoffry.robe...@gmail.comwrote: All, I have read about blueprint. I have reviewed the schema for blueprint, I have the specification. I did not find any mention of the use of elements like camelContext... or broker... or whatever. Yet I see these elements being used in blueprint files with success. My question is, Is there any documentation that catalogs all this? What other elements are there? Thanks -- Geoffry Roberts -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: EBA and Karaf - Can I install an EBA to KARAF
Aries already provides a deployer for eba files iirc. On Thu, Sep 29, 2011 at 07:29, Jean-Baptiste Onofré j...@nanthrax.net wrote: Hi, the EBA is an artifact format coming from Apache Aries. Currently, we don't have a deployer (polling the deploy folder) for this artifact. Anyway, it would be good to provide one. I will raise a Jira in that way. Regards JB On 09/29/2011 02:07 AM, Matt Madhavan wrote: Oh BTW, I copied the the eba to the deploy folder as well! Noting happens and no log neither -- View this message in context: http://karaf.922171.n3.nabble.** com/EBA-and-Karaf-Can-I-**install-an-EBA-to-KARAF-** tp3377768p339.htmlhttp://karaf.922171.n3.nabble.com/EBA-and-Karaf-Can-I-install-an-EBA-to-KARAF-tp3377768p339.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: EBA and Karaf - Can I install an EBA to KARAF
Yeah, that would definitely make sense. The bundle is the application-install one afaik: http://repo1.maven.org/maven2/org/apache/aries/application/org.apache.aries.application.install/0.3/ On Thu, Sep 29, 2011 at 07:49, Jean-Baptiste Onofré j...@nanthrax.net wrote: Thanks for the update Guillaume. Anyway, regarding the application-without-isolation feature, I can't see any bundle looking like a deployer. Maybe we would have just to add the deployer bundle in the feature. I will dig around that. Regards JB On 09/29/2011 07:46 AM, Guillaume Nodet wrote: Aries already provides a deployer for eba files iirc. On Thu, Sep 29, 2011 at 07:29, Jean-Baptiste Onofré j...@nanthrax.net mailto:j...@nanthrax.net wrote: Hi, the EBA is an artifact format coming from Apache Aries. Currently, we don't have a deployer (polling the deploy folder) for this artifact. Anyway, it would be good to provide one. I will raise a Jira in that way. Regards JB On 09/29/2011 02:07 AM, Matt Madhavan wrote: Oh BTW, I copied the the eba to the deploy folder as well! Noting happens and no log neither -- View this message in context: http://karaf.922171.n3.nabble.**__com/EBA-and-Karaf-Can-I-__** install-an-EBA-to-KARAF-__**tp3377768p339.html http://karaf.922171.n3.**nabble.com/EBA-and-Karaf-Can-** I-install-an-EBA-to-KARAF-**tp3377768p339.htmlhttp://karaf.922171.n3.nabble.com/EBA-and-Karaf-Can-I-install-an-EBA-to-KARAF-tp3377768p339.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Jean-Baptiste Onofré jbono...@apache.org mailto:jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: bundles command in 2.2.3
It's not really a command afaik. bundles was resolved to calling getBundles on the bundle context. You can now do $.context bundles if you want to access the list of deployed bundles. On Fri, Sep 30, 2011 at 10:03, Achim Nierbeck bcanh...@googlemail.comwrote: Could it be that this command is only available with equinox and not with felix? just something that crossed my mind with the possibility of verification :) 2011/9/29 Brad Beck brad.b...@quantumretail.com: bundles returned the actual bundle objects which could then be interrogated, which is what I really want... On Sep 29, 2011, at 3:49 PM, Glen Mazza wrote: osgi:list or list should simply do. From: Brad Beck [brad.b...@quantumretail.com] Sent: Wednesday, September 28, 2011 5:30 PM To: user@karaf.apache.org Subject: bundles command in 2.2.3 One used to be able to get a list of bundles at the console using bundles, the 2.2.3 manual even still references it. I can't seem to get it to work under 2.2.3. Was this an intentional change? If so, is there an alternative? Thanks, -Brad -- -- *Achim Nierbeck* Apache Karaf http://karaf.apache.org/ Committer PMC OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/ Committer Project Lead blog http://notizblog.nierbeck.de/ -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Add another Jetty Server / OSGI
I think pax-web uses factory configurations, so multiple configurations would lead to multiple http servers. On Thu, Oct 6, 2011 at 12:02, Jean-Baptiste Onofré j...@nanthrax.net wrote: OK, got it. Yes it's possible for the Jetty, but I'm not sure for the OSGi HTTP service. Let me check. Regards JB On 10/06/2011 11:56 AM, Charles Moulliard wrote: My idea is to have 2 separate Jetty Servers or one server with by example 2 connectors (8181, 8282) to be able to separate administration (webconsole, karaf console, ) from camel-cxf, camel-jetty that we use in applications and define different level of security, logging, On Thu, Oct 6, 2011 at 11:49 AM, Jean-Baptiste Onofréj...@nanthrax.net wrote: Hi Charles, Do you really need another Jetty ? I think just a new Jetty connector is enough. You can bind a port number for the HTTP service and another port for CXF for instance. WDYT ? Regards JB On 10/06/2011 11:46 AM, Charles Moulliard wrote: Hi, Is it possible to add a new Jetty Server (different from the one provided by default when installing features http or webconsole on Karaf) and register it as HTTP OSGI Service to allow by example CXF WS to be registered within the Servlet Container of this HTTP Server ? Is it something that we can do in blueprint / spring DM or a bundle must be created for that (java + xml config files) ? Regards, Charles Moulliard Apache Committer Blog : http://cmoulliard.blogspot.com Twitter : http://twitter.com/cmoulliard Linkedin : http://www.linkedin.com/in/**charlesmoulliardhttp://www.linkedin.com/in/charlesmoulliard Skype: cmoulliard -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Running Karaf without PAX Logging
The main problem I see with logback is that there's no properties based configuration for logback. That's a real problem imho because we would not be able to leverage the OSGi ConfigAdmin which is a key point in order to have a clean interaction point with all configurations. If you are aware of anything that would look like a properties based configurator for logback, things may changed. That's said, I've still haven't heard any real reasons to swtich to logback. The version of log4j that is embedded in pax-logging has been enhanced to fix some over synchronization problem (leading to performance issues) and also to add the features that were needed (like a split appender, etc...). What need do you have for logback ? On Fri, Oct 14, 2011 at 12:31, samspy sam.spyc...@gmail.com wrote: Is there any ongoing effort for a Logback backend to Karaf? I can only find rather old messages in this regard. Thanks, Sam -- View this message in context: http://karaf.922171.n3.nabble.com/Running-Karaf-without-PAX-Logging-tp3177504p3421252.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Running Karaf without PAX Logging
Maybe we should include your appender by default in pax-logging ? It sounds quite useful to me. On Fri, Oct 14, 2011 at 12:51, Achim Nierbeck bcanh...@googlemail.comwrote: Well, the only thing i would like to have would be the appender that already does the zipping of roled over log files. But this can be easily done with a custom appender and I wrote a blog [1] about it how to attach it to karaf, so no real need to actually get it into pax-logging ;) regards, Achim [1] - http://nierbeck.de/cgi-bin/weblog_basic/index.php?p=201 2011/10/14 Guillaume Nodet gno...@gmail.com The main problem I see with logback is that there's no properties based configuration for logback. That's a real problem imho because we would not be able to leverage the OSGi ConfigAdmin which is a key point in order to have a clean interaction point with all configurations. If you are aware of anything that would look like a properties based configurator for logback, things may changed. That's said, I've still haven't heard any real reasons to swtich to logback. The version of log4j that is embedded in pax-logging has been enhanced to fix some over synchronization problem (leading to performance issues) and also to add the features that were needed (like a split appender, etc...). What need do you have for logback ? On Fri, Oct 14, 2011 at 12:31, samspy sam.spyc...@gmail.com wrote: Is there any ongoing effort for a Logback backend to Karaf? I can only find rather old messages in this regard. Thanks, Sam -- View this message in context: http://karaf.922171.n3.nabble.com/Running-Karaf-without-PAX-Logging-tp3177504p3421252.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- -- *Achim Nierbeck* Apache Karaf http://karaf.apache.org/ Committer PMC OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/ Committer Project Lead blog http://notizblog.nierbeck.de/ -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Running Karaf without PAX Logging
On Fri, Oct 14, 2011 at 17:21, ceki c...@qos.ch wrote: On Oct 14, 2011 3:47:54 am Guillaume Nodet wrote: The main problem I see with logback is that there's no properties based configuration for logback. That's a real problem imho because we would not be able to leverage the OSGi ConfigAdmin which is a key point in order to have a clean interaction point with all configurations. If you are aware of anything that would look like a properties based configurator for logback, things may changed. Hello Guillaume, I am puzzled by the your remark. Logback can be configured via configuration files in XML or groovy format. It can also be configured programmatically via an API. (XML/groovy configuration invoke this API underneath). Logback can also be interrogated programmatically. For example, you can get obtain the list of existing logger and get (or set) the level of any logger. Why on earth would you care that logback does not support configuration files in properties format? What am I missing? I do care a lot about logback being configure with properties to be able to leverage ConfigAdmin. It should be *the* way to configure things in OSGi. That way, you can distribute the configuration remotely or store it in a DB or in any other mean without having to rewrite all the bundles to leverage that. That's the benefit of using a standard service. You just said the configuration file needs to be xml or groovy, which is different from a properties file. For config admin, the input data needs to be a map of key/value pairs. I haven't said it was not possible with logback, just that it does not exist, and I don't have the time and will to start writing a new configuration mechanism for logback without having any real need to switch to it. But if you want to try that, it could be nice. Though I still haven't heard the reasons why you want logback instead of pax-logging. Best regards, -- Ceki -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Running Karaf without PAX Logging
Well, if you really want to go that way, let's do it: - faster implementation: i explained we fixed most synchronization issues which were performance bottlenecks, so I'm not really convinced by that argument - extensive battery of tests: well, log4j is 10 years old and quite stable ;-) - loback speaks slf4j: in our case, that does not help, as we'd have to use a intermediate api anyway because of the other constraints of pax-logging - extensive documentation: maybe that's true, i haven't had many complains about that - configuration files in xml or groovy: we need key/value pairs, so we can't leverage those - automatic relaoding of config files: that's provided by config admin and we would not use it - lilith: i think it can be used anyway - conditional processing: we can't use xml or groovy - filters: we support MDC in pax-logging, i'm sure we can implement that if needed - sifting appender: it has been added in pax-logging - automatic compression of log files: Achim provided the code on github and i think we can easily add it to pax-logging - stack traces: lol, this actually comes from James Strachan who provided a patch to log4j and we have the same for osgi - automic removal: that's a minor feature we could add too In summary: most of those reasons may actually be true for log4j, but not for pax-logging. I don't have any problems with logback, it's just that I don't want to spend too much time to integrate it into pax-logging. You're very welcome to do it if you want, as long as we keep the same feature set. If you really need something in the above list that has not been fixed yet, you can either fix the problem in pax-logging / log4j, or enhance pax-logging to change the backend, but I just think the costs are very different. To your other reply, you're right, I want to enforce ConfigAdmin. On Fri, Oct 14, 2011 at 18:04, samspy sam.spyc...@gmail.com wrote: Hi Guillaume I assume you missed my http://karaf.922171.n3.nabble.com/Running-Karaf-without-PAX-Logging-tp3177504p3421434.html post , where I linked the published reasons for switching to logback as well as stating our own. Could you perhaps briefly explain why the backend logging configuration file needs to be read by non-backend karaf components? Thanks again, Sam -- View this message in context: http://karaf.922171.n3.nabble.com/Running-Karaf-without-PAX-Logging-tp3177504p3422108.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Running Karaf without PAX Logging
On Fri, Oct 14, 2011 at 18:27, ceki c...@qos.ch wrote: On 14/10/2011 5:48 PM, Guillaume Nodet wrote: Thank you for your quick response. I do care a lot about logback being configure with properties to be able to leverage ConfigAdmin. It should be *the* way to configure things in OSGi. That way, you can distribute the configuration remotely or store it in a DB or in any other mean without having to rewrite all the bundles to leverage that. That's the benefit of using a standard service. How does any of the above change for properties format. For log4j which supports properties format, you still need to invoke PropertyConfigurator on the properties (or some URL containing the properties). It would be no different with logback, except that you would invoke a different configurator. Yeah, I agree. I just don't want to write that configurator. You just said the configuration file needs to be xml or groovy, which is different from a properties file. For config admin, the input data needs to be a map of key/value pairs. I haven't said it was not possible with logback, just that it does not exist, and I don't have the time and will to start writing a new configuration mechanism for logback without having any real need to switch to it. But if you want to try that, it could be nice. Though I still haven't heard the reasons why you want logback instead of pax-logging. I am not a Karaf user, at least not yet. I am the founder of both log4j and logback projects although I now work mostly on logback. I am just trying to understand your use case for properties configuration. My apologies if the use case is obvious for Karaf users. The use case is to provide the configuration through the standard OSGi ConfigAdmin service. The main benefit is that the data storage can be abstracted. We use files by default as the primary source, but Cellar can push configuration changes between instances, Fuse Fabric stores them in Zookeeper, I've also seen people using a JDBC storage for the configuration. All the things are mostly interesting when dealing with large deployments, not for a single instance, I agree. Best regards, -- Ceki -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Running Karaf without PAX Logging
Yeah, but you loose the ability to easily configure one logger level or such. Fine grained configuration is much easier imho, but you're right, it would work too. On Fri, Oct 14, 2011 at 19:28, David Jencks david_jen...@yahoo.com wrote: It might not fit too well with the felix/karaf idea of installing config admin pid configs through .cfg property files, but config admin has no problem dealing with a string property that is an entire xml document. So I'd go for a handy way to initialize such config admin properties rather than a whole new logback configurator. Maybe some kind of notation like ..file.key=filelocation which means the value for key is read from the filelocation. thanks david jencks On Oct 14, 2011, at 9:33 AM, Guillaume Nodet wrote: On Fri, Oct 14, 2011 at 18:27, ceki c...@qos.ch wrote: On 14/10/2011 5:48 PM, Guillaume Nodet wrote: Thank you for your quick response. I do care a lot about logback being configure with properties to be able to leverage ConfigAdmin. It should be *the* way to configure things in OSGi. That way, you can distribute the configuration remotely or store it in a DB or in any other mean without having to rewrite all the bundles to leverage that. That's the benefit of using a standard service. How does any of the above change for properties format. For log4j which supports properties format, you still need to invoke PropertyConfigurator on the properties (or some URL containing the properties). It would be no different with logback, except that you would invoke a different configurator. Yeah, I agree. I just don't want to write that configurator. You just said the configuration file needs to be xml or groovy, which is different from a properties file. For config admin, the input data needs to be a map of key/value pairs. I haven't said it was not possible with logback, just that it does not exist, and I don't have the time and will to start writing a new configuration mechanism for logback without having any real need to switch to it. But if you want to try that, it could be nice. Though I still haven't heard the reasons why you want logback instead of pax-logging. I am not a Karaf user, at least not yet. I am the founder of both log4j and logback projects although I now work mostly on logback. I am just trying to understand your use case for properties configuration. My apologies if the use case is obvious for Karaf users. The use case is to provide the configuration through the standard OSGi ConfigAdmin service. The main benefit is that the data storage can be abstracted. We use files by default as the primary source, but Cellar can push configuration changes between instances, Fuse Fabric stores them in Zookeeper, I've also seen people using a JDBC storage for the configuration. All the things are mostly interesting when dealing with large deployments, not for a single instance, I agree. Best regards, -- Ceki -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Simple way to log all bundle events
I think you have to. The OSGi framework is not mandated to log bundle / service events, and Felix Framework does not, so you could either use a BundleListener or an EventHandler and log the events yourself. If you already use Camel, you could also look at http://gnodet.blogspot.com/2010/09/two-karaf-related-camel-components.htmland use the event component to grab osgi events and log them using the log component. On Mon, Oct 24, 2011 at 10:20, Zhemzhitsky Sergey sergey_zhemzhit...@troika.ru wrote: Hi there, ** ** I have to log all bundle events (started, starting, etc.) among with some bundle properties, so I’m wondering in whether there is a simple way to do it without developing of custom BundleListener? ** ** ** ** Best Regards, Sergey ___ The information contained in this message may be privileged and conf idential and protected from disclosure. If you are not the original intended recipient, you are hereby notified that any review, retransmission, dissemination, or other use of, or taking of any action in reliance upon, this information is prohibited. If you have received this communication in error, please notify the sender immediately by replying to this message and delete it from your computer. Thank you for your cooperation. Troika Dialog, Russia. If you need assistance please contact our Contact Center (+7495) 258 0500or go to www.troika.ru/eng/Contacts/system.wbp -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Is it possible to authorize a local karaf client without providing login information?
Fwiw, we'd have to also enhance our ssh server integration to support certificate based authentication. The sshd code itself can support that, but we'd have to implement an org.apache.sshd.server.PublickeyAuthenticator and provide the needed means to manage the certificates in Karaf. On Tue, Oct 25, 2011 at 10:50, Jean-Baptiste Onofré j...@nanthrax.net wrote: Correct, the release is not available, and no, it's not yet scheduled. Feel free to create the Jira (or I will do it if you want). Regards JB On 10/25/2011 10:45 AM, mst wrote: Hi JB, thanks for the quick response. I guess ...we can enhance the client... means that this feature is not available right now. Is it scheduled for a certain release? Should I create an issue? Best Markus -- View this message in context: http://karaf.922171.n3.nabble.** com/Is-it-possible-to-**authorize-a-local-karaf-** client-without-providing-**login-information-**tp3450816p3450846.htmlhttp://karaf.922171.n3.nabble.com/Is-it-possible-to-authorize-a-local-karaf-client-without-providing-login-information-tp3450816p3450846.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Aries 0.1 - Latest Karaf version I can use
You could try with karaf 2.0.0, as 2.1.0 is using Aries 0.2. Though I don't really remember if 2.1 could work with aries 0.1. You try, you could download a karaf version, launch it and update the aries bundles with the 0.1 version to see if it still works. On Thu, Oct 27, 2011 at 21:23, Matt Madhavan mattmadha...@gmail.com wrote: Hello, Due to my client's target OSgG environment (IBM WAS 7.1) I;m stuck with Aries 0.1. I'm now trying to create a target environment with KARAF and Aries 0.1 bundles from IBM! Can some one let me know whats the latest Karaf version I can use that will work with Aries 0.1? Also I would like to have Camel installed as well. Thanks in Advance! Matt Madhavan -- View this message in context: http://karaf.922171.n3.nabble.com/Aries-0-1-Latest-Karaf-version-I-can-use-tp3458809p3458809.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Aries 0.1 - Latest Karaf version I can use
Camel is not really dependant on Karaf, so any version should work. However, camel-blueprint will have some requirement on aries (so may not be able to use that part). The commands that have been added recently will of course not work either. On Thu, Oct 27, 2011 at 22:32, Matt Madhavan mattmadha...@gmail.com wrote: Hi Guillaume, I could make it work with 2.0.0 but now Camel does not like Karaf 2.0.0. Any idea which version of Camel will work with Karaf 2.0.0? Man IBM! Thanks Matt -- View this message in context: http://karaf.922171.n3.nabble.com/Aries-0-1-Latest-Karaf-version-I-can-use-tp3458809p3459051.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Aries and Spring Co-Existance in Karaf
You can use OSGi services for that. OSGi services can be exported and imported irrespective of the underlying technology used. On Tue, Nov 1, 2011 at 13:35, Raman Gupta rocketra...@gmail.com wrote: On 11/01/2011 06:05 AM, Ioannis Canellos wrote: Let's not confuse blueprint with spring. Blueprint is a declarative way to work with OSGi services and Spring is a framework for creating applications. I don't think that Aries has the same focus with Spring but with SpringDM. You can always use both, if you have to go with Spring. If I had to use Spring, I would use it only where its necessary and for managing services etc I would use Aries. Example: In Cellar 90% of the modules use Aries, but there is a single module that uses Spring/SpringDM. We don't have any problem with that. What would have been nice is if Blueprint provided a way, out of the box, to expose beans created by Spring or Guice to the Blueprint context. That way, one could use the DI framework of choice / annotations inside a bundle, while consistently using Blueprint as a microservice layer. I'm surprised the Blueprint spec developers didn't consider interop with existing DI frameworks as a first class spec item. I suppose such functionality could still be implemented as a Blueprint extension for each DI framework. Regards, Raman Gupta VIVO Systems http://vivosys.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: shell commands user roles
Not really, while that could be implemented for commands, the problem is that the command line also allows introspection and scripting and authorization can't easily be done at that level so the console would not be totally secured anyway. On Wed, Nov 2, 2011 at 16:25, rrsavage rrsav...@hotmail.com wrote: I'm new to Karaf and have a question about user access control within the (SSH) shell. Is there a way to define more granular level of user access to see (list/autocomplete) and execute commands via the (SSH) shell? For example, can certain commands be restricted to a configured set of user roles via the command's name or scope? Thanks, Robert -- View this message in context: http://karaf.922171.n3.nabble.com/shell-commands-user-roles-tp3474148p3474148.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: shell commands user roles
We'd have to keep anything we do on role based access consistent with the web console / jmx management layers. On Wed, Nov 2, 2011 at 17:16, Achim Nierbeck bcanh...@googlemail.comwrote: Hi JB, Robert sounds like a reasonable User/Role feature for Karaf, would be interesting to see what roles we have here, a full accessible admin, a user level, do we need more distinct levels, like for example features, web, that correspond to the std. feature sets we have? @Robert could you open a Jira issue for that feature request :) regards, Achim 2011/11/2 Jean-Baptiste Onofré j...@nanthrax.net Hi Robert, it's not possible for now but it's a good idea. We have something similar in Apache Kalumet (called AccessList). It's a good new feature for Karaf 3.0. Regards JB On 11/02/2011 04:58 PM, rrsavage wrote: Really what I'm after is a two level access system. An admin level that has full access to all commands, scripting, introspection, etc. And a user level of access that perhaps only provides access to a limited number of command. Additionally user level access would disallow scripting and introspection capabilities. Is this a reasonable approach and is it even possible? Thanks, Robert -- View this message in context: http://karaf.922171.n3.nabble.** com/shell-commands-user-roles-**tp3474148p3474241.htmlhttp://karaf.922171.n3.nabble.com/shell-commands-user-roles-tp3474148p3474241.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- -- *Achim Nierbeck* Apache Karaf http://karaf.apache.org/ Committer PMC OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/ Committer Project Lead blog http://notizblog.nierbeck.de/ -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Seamlessly switching from Felix and Equinox - data folder
That's not really possible because each framework stores its data in a proprietary format. It's not something I would advise to do anyway, unless you're still in development mode in which case bou may want to take a look at pax-exam. On Thu, Nov 3, 2011 at 22:15, Matt Madhavan mattmadha...@gmail.com wrote: Hello, My understanding is that when I switch OSGi runtime the data folder becomes stale and I have to start delete the data folder content and start installing my features and bundles again. Can I preserve the state of my Karaf between OSGi runtime switches? Will save some time! Thanks Matt -- View this message in context: http://karaf.922171.n3.nabble.com/Seamlessly-switching-from-Felix-and-Equinox-data-folder-tp3478439p3478439.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Fetching feature from a nexus repository
I think in some cases, the error message can be misleading. Can you check that the jar returned by your nexus is a real jar and not an error page? On Tue, Nov 15, 2011 at 13:02, Michael Prieß mailingliste...@googlemail.com wrote: Hello, my Karaf Installation is behind a Firewall, so I have to use my own Maven Repository in the same subnet (Nexus) to fetch the wrapper feature from my nexus. So I changed the file org.ops4j.pax.url.mvn.cfg, uncommented the other repository and added my repository: org.ops4j.pax.url.mvn.repositories= \ http://myRepro/content/groups/public/ If I now run a feature:install wrapper I get the following Error: Manifest not present in the first entry of the zip mvn:org.apache.karaf.shell/org.apache.karaf.shell.wrapper/2.2.4 Have anyone a idea how fix this failure? Regards, Michael -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: OSGI Fragment and Logging
You don't need to import that package because your fragment will be associated to the pax-logging-service bundle which contain that class, so it will be directly available. On Wed, Nov 16, 2011 at 15:27, Hervé BARRAULT herve.barra...@gmail.comwrote: Hi Achim My problem is the fragment is not able to be resolved as the package org.apache.log4j.helpers can't be imported. I thought Pax Logging export it but it does not. Adding again all the log4j API is not i think the best idea. Regards Hervé On 11/16/11, Achim Nierbeck bcanh...@googlemail.com wrote: Hi Hervé thanx for the hint. Did you also take a look a the configuring part? Cause to get things going in Karaf/ServiceMix you need to make sure your fragment is available before the host bundle enters the resolved state. regards, Achim 2011/11/16 Hervé BARRAULT herve.barra...@gmail.com Thanks for the answer. I have seen your blog but my case is more a problem of dependency management than building a fragment itself. PS: be careful, in your blog, close and open xml tag don't match : import -Package!*/import fragment -Hostorg.ops4j.pax.logging.pax-logging-service/fragment On 11/16/11, Achim Nierbeck bcanh...@googlemail.com wrote: Hi I wrote a blog about it [1] regards, Achim [1] - http://nierbeck.de/cgi-bin/weblog_basic/index.php?p=201 2011/11/16 Hervé BARRAULT herve.barra...@gmail.com Hi, I am using using ServiceMix 4.3.0-fuse-01-00 so Karaf http://fusesource.com/wiki/display/ProdInfo/Fuse+Karaf+v2.0.0+Release+Notes 2.0.0-fuse-00-00 I have seen the following documentation : http://karaf.apache.org/manual/2.2.2/users-guide/logging-system.html Especially, the Using your own appenders section. I have tried to do something like the documentation but i have an issue. For my application i need to use the CountingQuietWriter class which is in the org.apache.log4j.helpers package. But it seems that pax-logging-api does not export this package. What is the best way to import this package ? For information (i know it is an old version) but pax-logging-api [1.5.2] export log4j 1.2.15 packages (looking to the export-package) and pax-logging-service [1.5.2] embeds a log4j 1.2.16 (looking to the pom) Thanks For Answers Hervé -- *Achim Nierbeck* Apache Karaf http://karaf.apache.org/ Committer PMC OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/ Committer Project Lead blog http://notizblog.nierbeck.de/ -- *Achim Nierbeck* Apache Karaf http://karaf.apache.org/ Committer PMC OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/ Committer Project Lead blog http://notizblog.nierbeck.de/ -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: OSGI Fragment and Logging
Try adding Import-Package!*/Import-Package in your pom On Wed, Nov 16, 2011 at 15:41, Hervé BARRAULT herve.barra...@gmail.comwrote: Hi Guillaume, I am using the maven-bundle-plugin (1.4.0), and i don't find a way to build a bundle without importing all needed packages. Regards Hervé On 11/16/11, Guillaume Nodet gno...@gmail.com wrote: You don't need to import that package because your fragment will be associated to the pax-logging-service bundle which contain that class, so it will be directly available. On Wed, Nov 16, 2011 at 15:27, Hervé BARRAULT herve.barra...@gmail.comwrote: Hi Achim My problem is the fragment is not able to be resolved as the package org.apache.log4j.helpers can't be imported. I thought Pax Logging export it but it does not. Adding again all the log4j API is not i think the best idea. Regards Hervé On 11/16/11, Achim Nierbeck bcanh...@googlemail.com wrote: Hi Hervé thanx for the hint. Did you also take a look a the configuring part? Cause to get things going in Karaf/ServiceMix you need to make sure your fragment is available before the host bundle enters the resolved state. regards, Achim 2011/11/16 Hervé BARRAULT herve.barra...@gmail.com Thanks for the answer. I have seen your blog but my case is more a problem of dependency management than building a fragment itself. PS: be careful, in your blog, close and open xml tag don't match : import -Package!*/import fragment -Hostorg.ops4j.pax.logging.pax-logging-service/fragment On 11/16/11, Achim Nierbeck bcanh...@googlemail.com wrote: Hi I wrote a blog about it [1] regards, Achim [1] - http://nierbeck.de/cgi-bin/weblog_basic/index.php?p=201 2011/11/16 Hervé BARRAULT herve.barra...@gmail.com Hi, I am using using ServiceMix 4.3.0-fuse-01-00 so Karaf http://fusesource.com/wiki/display/ProdInfo/Fuse+Karaf+v2.0.0+Release+Notes 2.0.0-fuse-00-00 I have seen the following documentation : http://karaf.apache.org/manual/2.2.2/users-guide/logging-system.html Especially, the Using your own appenders section. I have tried to do something like the documentation but i have an issue. For my application i need to use the CountingQuietWriter class which is in the org.apache.log4j.helpers package. But it seems that pax-logging-api does not export this package. What is the best way to import this package ? For information (i know it is an old version) but pax-logging-api [1.5.2] export log4j 1.2.15 packages (looking to the export-package) and pax-logging-service [1.5.2] embeds a log4j 1.2.16 (looking to the pom) Thanks For Answers Hervé -- *Achim Nierbeck* Apache Karaf http://karaf.apache.org/ Committer PMC OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/ Committer Project Lead blog http://notizblog.nierbeck.de/ -- *Achim Nierbeck* Apache Karaf http://karaf.apache.org/ Committer PMC OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/ Committer Project Lead blog http://notizblog.nierbeck.de/ -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Bouncy Castle JCE with Felix
It definitely looks like a bug in Felix. Please raise a JIRA for it. On Thu, Dec 1, 2011 at 17:08, Caspar MacRae ear...@gmail.com wrote: Hi, I've an issue with Bouncy Castle JCE running with Felix (I found this while trying to extend our custom Karaf distro, it seems to be a problem with Karaf 2.2.0 (Felix 3.0.8) through to 2.2.4 (Felix 3.0.9) but doesn't happen when I use Equinox. Could not create framework: java.lang.ArrayIndexOutOfBoundsException: -1 java.lang.ArrayIndexOutOfBoundsException: -1 at java.util.ArrayList.get(ArrayList.java:324) at org.apache.felix.framework.BundleImpl.getCurrentModule(BundleImpl.java:1050) at org.apache.felix.framework.BundleImpl.getSymbolicName(BundleImpl.java:859) at org.apache.felix.framework.Felix.toString(Felix.java:1019) at org.apache.felix.framework.Logger.doLog(Logger.java:128) at org.apache.felix.framework.Logger._log(Logger.java:181) at org.apache.felix.framework.Logger.log(Logger.java:114) at org.apache.felix.framework.ExtensionManager.init(ExtensionManager.java:201) at org.apache.felix.framework.Felix.init(Felix.java:374) at org.apache.felix.framework.FrameworkFactory.newFramework(FrameworkFactory.java:28) at org.apache.karaf.main.Main.launch(Main.java:266) at org.apache.karaf.main.Main.main(Main.java:427) Steps to reproduce: wget http://www.apache.org/dyn/closer.cgi/karaf/2.2.4/apache-karaf-2.2.4.tar.gz tar -xvzf apache-karaf-2.2.4.tar.gz cd apache-karaf-2.2.4/ # Assuming you've got bcprov-jdk16-1.46.jar in your maven repo cp ~/.m2/repository/org/bouncycastle/bcprov-jdk16/1.46/bcprov-jdk16-1.46.jar ./lib/ext/ nano etc/custom.properties # Add the following: org.osgi.framework.system.packages.extra = \ org.bouncycastle.math.ec;version=1.46; \ org.bouncycastle.jce.provider;version=1.46; org.apache.felix.karaf.security.providers = org.bouncycastle.jce.provider.BouncyCastleProvider org.osgi.framework.bootdelegation = org.bouncycastle.*; ./bin/karaf # It exits immediately with the stacktrace above Am I doing something incredibly stupid or should a bug be raised (with Karaf or Felix)? thanks, Caspar -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ Open Source SOA http://fusesource.com
Re: Karaf - managing Camel routes startup behavior
You should listen to CamelContext objects being registered as OSGi services.
If they all have autostart = false, you could then manually control them.
On Mon, Jan 23, 2012 at 10:51, kmoens kris_mo...@yahoo.com wrote:
Hi JB,
thanks for the idea. Not sure this covers our needs, we want to control
automatically the startup of deployed routes at both boottime of karaf and
deploy time of a new bundle containing routes.
In this case the bean you proposed should be listening on 'deploy events' of
bundles having routes in spring xml files, how to achieve that?
When I activate trace, I see the follwing messages logged:
13:41:18,915 | INFO | SpringOsgiExtenderThread-3 |
OsgiBundleXmlApplicationContext | 47 - org.springframework.context -
3.0.5.RELEASE | Refreshing
OsgiBundleXmlApplicationContext(bundle=com.my.samples.Sample1,
config=osgibundle:/META-INF/spring/*.xml): startup date [Wed Jan 18 13:41:18
CET 2012]; root of context hierarchy
13:41:18,915 | TRACE | SpringOsgiExtenderThread-3 |
OsgiBundleResourcePatternResolver | 60 - org.springframework.osgi.io - 1.2.1
| Found root resources for [osgibundle:/META-INF/spring/] :{URL
[bundle://138.0:0/META-INF/spring/]}
13:41:18,915 | TRACE | SpringOsgiExtenderThread-3 |
OsgiBundleResourcePatternResolver | 60 - org.springframework.osgi.io - 1.2.1
| Resolved location pattern [osgibundle:/META-INF/spring/*.xml] to resources
[URL [bundle://138.0:0/META-INF/spring/Sample1.xml]]
I was wondering if we could find a way to extend the spring deploy process
some how and add our condition.
BRs,
Kris
--
View this message in context:
http://karaf.922171.n3.nabble.com/Karaf-managing-Camel-routes-startup-behavior-tp3669403p3681405.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
Re: Logging using log4j filters
The filter support has been added in pax-logging.
Have a look at
https://github.com/ops4j/org.ops4j.pax.logging/blob/master/pax-logging-service/src/main/java/org/apache/log4j/PaxLoggingConfigurator.java
You may very well be right that the order isn't kept, which would
definitely be a bug.
On Mon, Jan 30, 2012 at 10:17, Bengt Rodehav be...@rodehav.com wrote:
I have the following configuration in my org.ops4j.pax.logging.cfg:
# Per bundle log at INFO level
log4j.appender.bundle=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle.key=bundle.name
log4j.appender.bundle.default=karaf
log4j.appender.bundle.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle.appender.MaxFileSize=10MB
log4j.appender.bundle.appender.MaxBackupIndex=2
log4j.appender.bundle.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle.appender.layout.ConversionPattern=%d{ISO8601} | %-5.5p
| %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle.appender.file=${logdir}/bundles/$\\{bundle.name\\}.log
log4j.appender.bundle.appender.append=true
log4j.appender.bundle.threshold=INFO
# TRACE level for specific bundle - should normally be disabled
log4j.appender.bundle_trace=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle_trace.key=bundle.name
log4j.appender.bundle_trace.default=karaf
log4j.appender.bundle_trace.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle_trace.appender.MaxFileSize=20MB
log4j.appender.bundle_trace.appender.MaxBackupIndex=1
log4j.appender.bundle_trace.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle_trace.appender.layout.ConversionPattern=%d{ISO8601} |
%-5.5p | %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle_trace.appender.file=${logdir}/bundles/trace/$\\{bundle.name\\}.log
log4j.appender.bundle_trace.appender.append=true
log4j.appender.bundle_trace.threshold=TRACE
log4j.appender.bundle_trace.filter.accept=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.accept.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.accept.AcceptOnMatch=false
log4j.appender.bundle_trace.filter.deny=org.apache.log4j.varia.DenyAllFilter
The intention is to have bundle specific logs at INFO level but have a
separate TRACE log for a specific bundle. The latter is not enabled by
default but only when debugging.
The problem is that the DenyAllFilter seems to take precedence over the
StringMatchFilter. I believe that when listed in the order I do, the bundle
with the name org.apache.camel.camel-core should be logged at TRACE level
but no other bundles. Could it be that the ordering of filters are not
preserved? I think that native log4j only supports filters when using XML
configuration and I assume that the Karaf filtering support has been added
on top of log4j (or is it in Pax-logging)? Has the ordering of filters been
taken into account?
I've been testing this on Karaf 2.2.0 with Pax logging 1.6.0.
/Bengt
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
Re: Logging using log4j filters
Actually, the filters support is built into log4j, but if there's
really a problem we can always fix it in pax-logging until the patch
is released in log4j.
On Mon, Jan 30, 2012 at 10:21, Guillaume Nodet gno...@gmail.com wrote:
The filter support has been added in pax-logging.
Have a look at
https://github.com/ops4j/org.ops4j.pax.logging/blob/master/pax-logging-service/src/main/java/org/apache/log4j/PaxLoggingConfigurator.java
You may very well be right that the order isn't kept, which would
definitely be a bug.
On Mon, Jan 30, 2012 at 10:17, Bengt Rodehav be...@rodehav.com wrote:
I have the following configuration in my org.ops4j.pax.logging.cfg:
# Per bundle log at INFO level
log4j.appender.bundle=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle.key=bundle.name
log4j.appender.bundle.default=karaf
log4j.appender.bundle.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle.appender.MaxFileSize=10MB
log4j.appender.bundle.appender.MaxBackupIndex=2
log4j.appender.bundle.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle.appender.layout.ConversionPattern=%d{ISO8601} | %-5.5p
| %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle.appender.file=${logdir}/bundles/$\\{bundle.name\\}.log
log4j.appender.bundle.appender.append=true
log4j.appender.bundle.threshold=INFO
# TRACE level for specific bundle - should normally be disabled
log4j.appender.bundle_trace=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle_trace.key=bundle.name
log4j.appender.bundle_trace.default=karaf
log4j.appender.bundle_trace.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle_trace.appender.MaxFileSize=20MB
log4j.appender.bundle_trace.appender.MaxBackupIndex=1
log4j.appender.bundle_trace.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle_trace.appender.layout.ConversionPattern=%d{ISO8601} |
%-5.5p | %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle_trace.appender.file=${logdir}/bundles/trace/$\\{bundle.name\\}.log
log4j.appender.bundle_trace.appender.append=true
log4j.appender.bundle_trace.threshold=TRACE
log4j.appender.bundle_trace.filter.accept=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.accept.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.accept.AcceptOnMatch=false
log4j.appender.bundle_trace.filter.deny=org.apache.log4j.varia.DenyAllFilter
The intention is to have bundle specific logs at INFO level but have a
separate TRACE log for a specific bundle. The latter is not enabled by
default but only when debugging.
The problem is that the DenyAllFilter seems to take precedence over the
StringMatchFilter. I believe that when listed in the order I do, the bundle
with the name org.apache.camel.camel-core should be logged at TRACE level
but no other bundles. Could it be that the ordering of filters are not
preserved? I think that native log4j only supports filters when using XML
configuration and I assume that the Karaf filtering support has been added
on top of log4j (or is it in Pax-logging)? Has the ordering of filters been
taken into account?
I've been testing this on Karaf 2.2.0 with Pax logging 1.6.0.
/Bengt
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
Re: Logging using log4j filters
No, the support has been added in log4j:
http://svn.apache.org/viewvc?view=revisionrevision=821430
On Mon, Jan 30, 2012 at 10:30, Bengt Rodehav be...@rodehav.com wrote:
Hello Guillaume,
Doesn't the filter support in log4j require XML configuration (not
properties file)? If so, then I assume that Pax-logging has added the
possibility to use filters using a properties file configuration.
/Bengt
2012/1/30 Guillaume Nodet gno...@gmail.com
Actually, the filters support is built into log4j, but if there's
really a problem we can always fix it in pax-logging until the patch
is released in log4j.
On Mon, Jan 30, 2012 at 10:21, Guillaume Nodet gno...@gmail.com wrote:
The filter support has been added in pax-logging.
Have a look at
https://github.com/ops4j/org.ops4j.pax.logging/blob/master/pax-logging-service/src/main/java/org/apache/log4j/PaxLoggingConfigurator.java
You may very well be right that the order isn't kept, which would
definitely be a bug.
On Mon, Jan 30, 2012 at 10:17, Bengt Rodehav be...@rodehav.com wrote:
I have the following configuration in my org.ops4j.pax.logging.cfg:
# Per bundle log at INFO level
log4j.appender.bundle=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle.key=bundle.name
log4j.appender.bundle.default=karaf
log4j.appender.bundle.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle.appender.MaxFileSize=10MB
log4j.appender.bundle.appender.MaxBackupIndex=2
log4j.appender.bundle.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle.appender.layout.ConversionPattern=%d{ISO8601} |
%-5.5p
| %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle.appender.file=${logdir}/bundles/$\\{bundle.name\\}.log
log4j.appender.bundle.appender.append=true
log4j.appender.bundle.threshold=INFO
# TRACE level for specific bundle - should normally be disabled
log4j.appender.bundle_trace=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle_trace.key=bundle.name
log4j.appender.bundle_trace.default=karaf
log4j.appender.bundle_trace.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle_trace.appender.MaxFileSize=20MB
log4j.appender.bundle_trace.appender.MaxBackupIndex=1
log4j.appender.bundle_trace.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle_trace.appender.layout.ConversionPattern=%d{ISO8601}
|
%-5.5p | %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle_trace.appender.file=${logdir}/bundles/trace/$\\{bundle.name\\}.log
log4j.appender.bundle_trace.appender.append=true
log4j.appender.bundle_trace.threshold=TRACE
log4j.appender.bundle_trace.filter.accept=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.accept.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.accept.AcceptOnMatch=false
log4j.appender.bundle_trace.filter.deny=org.apache.log4j.varia.DenyAllFilter
The intention is to have bundle specific logs at INFO level but have a
separate TRACE log for a specific bundle. The latter is not enabled by
default but only when debugging.
The problem is that the DenyAllFilter seems to take precedence over the
StringMatchFilter. I believe that when listed in the order I do, the
bundle
with the name org.apache.camel.camel-core should be logged at TRACE
level
but no other bundles. Could it be that the ordering of filters are not
preserved? I think that native log4j only supports filters when using
XML
configuration and I assume that the Karaf filtering support has been
added
on top of log4j (or is it in Pax-logging)? Has the ordering of filters
been
taken into account?
I've been testing this on Karaf 2.2.0 with Pax logging 1.6.0.
/Bengt
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
Re: Logging using log4j filters
Looking at the log4j code, it seems the filters are ordered using
their ids, so in your case accept and deny.
So I think the order should be ok. Can you try changing their name so
that the order would be reversed ?
On Mon, Jan 30, 2012 at 11:09, Bengt Rodehav be...@rodehav.com wrote:
OK - I didn't know that.
Do you think I should post a message on ops4j's mailing list about this?
The reason I tried the Karaf mailing list first is that I believe this
whould be a pretty common (and useful) configuration. In my case, I will
probably create logs per camel context and not per bundle but I still need
the possiblity to configure more detailed logging for a specific MDC value.
Have you tried something similar yourself?
I actually posted a question on Stackoverflow about this as well:
http://stackoverflow.com/questions/9049119/set-log-level-based-on-mdc-value-in-log4j
No replies unfortunately. The filtering approach would be an alternative
(although not as elegant way) to accomplish what I wanted.
/Bengt
2012/1/30 Guillaume Nodet gno...@gmail.com
No, the support has been added in log4j:
http://svn.apache.org/viewvc?view=revisionrevision=821430
On Mon, Jan 30, 2012 at 10:30, Bengt Rodehav be...@rodehav.com wrote:
Hello Guillaume,
Doesn't the filter support in log4j require XML configuration (not
properties file)? If so, then I assume that Pax-logging has added the
possibility to use filters using a properties file configuration.
/Bengt
2012/1/30 Guillaume Nodet gno...@gmail.com
Actually, the filters support is built into log4j, but if there's
really a problem we can always fix it in pax-logging until the patch
is released in log4j.
On Mon, Jan 30, 2012 at 10:21, Guillaume Nodet gno...@gmail.com
wrote:
The filter support has been added in pax-logging.
Have a look at
https://github.com/ops4j/org.ops4j.pax.logging/blob/master/pax-logging-service/src/main/java/org/apache/log4j/PaxLoggingConfigurator.java
You may very well be right that the order isn't kept, which would
definitely be a bug.
On Mon, Jan 30, 2012 at 10:17, Bengt Rodehav be...@rodehav.com
wrote:
I have the following configuration in my org.ops4j.pax.logging.cfg:
# Per bundle log at INFO level
log4j.appender.bundle=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle.key=bundle.name
log4j.appender.bundle.default=karaf
log4j.appender.bundle.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle.appender.MaxFileSize=10MB
log4j.appender.bundle.appender.MaxBackupIndex=2
log4j.appender.bundle.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle.appender.layout.ConversionPattern=%d{ISO8601}
|
%-5.5p
| %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle.appender.file=${logdir}/bundles/$\\{bundle.name\\}.log
log4j.appender.bundle.appender.append=true
log4j.appender.bundle.threshold=INFO
# TRACE level for specific bundle - should normally be disabled
log4j.appender.bundle_trace=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle_trace.key=bundle.name
log4j.appender.bundle_trace.default=karaf
log4j.appender.bundle_trace.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle_trace.appender.MaxFileSize=20MB
log4j.appender.bundle_trace.appender.MaxBackupIndex=1
log4j.appender.bundle_trace.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle_trace.appender.layout.ConversionPattern=%d{ISO8601}
|
%-5.5p | %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle_trace.appender.file=${logdir}/bundles/trace/$\\{bundle.name\\}.log
log4j.appender.bundle_trace.appender.append=true
log4j.appender.bundle_trace.threshold=TRACE
log4j.appender.bundle_trace.filter.accept=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.accept.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.accept.AcceptOnMatch=false
log4j.appender.bundle_trace.filter.deny=org.apache.log4j.varia.DenyAllFilter
The intention is to have bundle specific logs at INFO level but have
a
separate TRACE log for a specific bundle. The latter is not enabled
by
default but only when debugging.
The problem is that the DenyAllFilter seems to take precedence over
the
StringMatchFilter. I believe that when listed in the order I do, the
bundle
with the name org.apache.camel.camel-core should be logged at
TRACE
level
but no other bundles. Could it be that the ordering of filters are
not
preserved? I think that native log4j only supports filters when
using
XML
configuration and I assume that the Karaf filtering support has been
added
on top of log4j (or is it in Pax-logging)? Has the ordering of
filters
been
taken into account?
I've been
Re: Logging using log4j filters
Have you tried matching on something more simple such as camel ?
The StringToMatch filter simply looks for the string to be in the
rendered event so maybe none of you events contains
bunde.name:org.apache.camel.camel-core
On Mon, Jan 30, 2012 at 11:52, Bengt Rodehav be...@rodehav.com wrote:
I tried these four combinations:
# 1
log4j.appender.bundle_trace.filter.a=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.a.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.a.AcceptOnMatch=true
log4j.appender.bundle_trace.filter.b=org.apache.log4j.varia.DenyAllFilter
# 2
log4j.appender.bundle_trace.filter.b=org.apache.log4j.varia.DenyAllFilter
log4j.appender.bundle_trace.filter.a=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.a.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.a.AcceptOnMatch=true
# 3
log4j.appender.bundle_trace.filter.a=org.apache.log4j.varia.DenyAllFilter
log4j.appender.bundle_trace.filter.b=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.b.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.b.AcceptOnMatch=true
# 4
log4j.appender.bundle_trace.filter.b=org.apache.log4j.varia.StringMatchFilter
log4j.appender.bundle_trace.filter.b.StringToMatch=bunde.name:org.apache.camel.camel-core
log4j.appender.bundle_trace.filter.b.AcceptOnMatch=true
log4j.appender.bundle_trace.filter.a=org.apache.log4j.varia.DenyAllFilter
This would check if ordering of the configurations or filter naming would
make a difference. Unfortunately none of the above work.
But as soon as I comment out the DenyAllFilter, trace logfiles appear in the
trace folder. So, either the DenyAllFilter prevents the StringMatchFilter
from working or the StringMatchFilter never matches...
/Bengt
2012/1/30 Guillaume Nodet gno...@gmail.com
Looking at the log4j code, it seems the filters are ordered using
their ids, so in your case accept and deny.
So I think the order should be ok. Can you try changing their name so
that the order would be reversed ?
On Mon, Jan 30, 2012 at 11:09, Bengt Rodehav be...@rodehav.com wrote:
OK - I didn't know that.
Do you think I should post a message on ops4j's mailing list about this?
The reason I tried the Karaf mailing list first is that I believe this
whould be a pretty common (and useful) configuration. In my case, I will
probably create logs per camel context and not per bundle but I still
need
the possiblity to configure more detailed logging for a specific MDC
value.
Have you tried something similar yourself?
I actually posted a question on Stackoverflow about this as well:
http://stackoverflow.com/questions/9049119/set-log-level-based-on-mdc-value-in-log4j
No replies unfortunately. The filtering approach would be an alternative
(although not as elegant way) to accomplish what I wanted.
/Bengt
2012/1/30 Guillaume Nodet gno...@gmail.com
No, the support has been added in log4j:
http://svn.apache.org/viewvc?view=revisionrevision=821430
On Mon, Jan 30, 2012 at 10:30, Bengt Rodehav be...@rodehav.com wrote:
Hello Guillaume,
Doesn't the filter support in log4j require XML configuration (not
properties file)? If so, then I assume that Pax-logging has added the
possibility to use filters using a properties file configuration.
/Bengt
2012/1/30 Guillaume Nodet gno...@gmail.com
Actually, the filters support is built into log4j, but if there's
really a problem we can always fix it in pax-logging until the patch
is released in log4j.
On Mon, Jan 30, 2012 at 10:21, Guillaume Nodet gno...@gmail.com
wrote:
The filter support has been added in pax-logging.
Have a look at
https://github.com/ops4j/org.ops4j.pax.logging/blob/master/pax-logging-service/src/main/java/org/apache/log4j/PaxLoggingConfigurator.java
You may very well be right that the order isn't kept, which would
definitely be a bug.
On Mon, Jan 30, 2012 at 10:17, Bengt Rodehav be...@rodehav.com
wrote:
I have the following configuration in
my org.ops4j.pax.logging.cfg:
# Per bundle log at INFO level
log4j.appender.bundle=org.apache.log4j.sift.MDCSiftingAppender
log4j.appender.bundle.key=bundle.name
log4j.appender.bundle.default=karaf
log4j.appender.bundle.appender=org.apache.log4j.RollingFileAppender
log4j.appender.bundle.appender.MaxFileSize=10MB
log4j.appender.bundle.appender.MaxBackupIndex=2
log4j.appender.bundle.appender.layout=org.apache.log4j.PatternLayout
log4j.appender.bundle.appender.layout.ConversionPattern=%d{ISO8601}
|
%-5.5p
| %-16.16t | %-32.32c{1} | %-32.32C %4L | %m%n
log4j.appender.bundle.appender.file=${logdir}/bundles/$\\{bundle.name\\}.log
Re: Apache Commons DBCP fragment
First question, why don't you use the existing bundles: http://repo1.maven.org/maven2/org/apache/servicemix/bundles/org.apache.servicemix.bundles.commons-dbcp/1.4_1/ For fragments, the import and export are added to the host, but I don't think this includes dynamic imports. On Mon, Jan 30, 2012 at 17:01, lbu lburgazz...@gmail.com wrote: Hi, I'm working to a project OSGi-fication and I have a little trouble with a fragment that does not seem taken into account by Karaf (2.2.5). The fragment is supposed to add DynamicImport-Package to org.apache.commons.dbcp to look-up jdbc drivers. Here the META-INF/MANIFEST.MF of my fragment: Manifest-Version: 1.0 Created-By: LB Bundle-Name: lb.fragment.apache.commons.dbcp Bundle-Vendor: LB Bundle-Version: 1.0.7 Bundle-SymbolicName: org.apache.commons.dbcp.fragment Bundle-Description: Fragment attached to Apache Commons DBCP Fragment-Host: org.apache.commons.dbcp DynamicImport-Package: * What's wrong ? Thx, Luca -- View this message in context: http://karaf.922171.n3.nabble.com/Apache-Commons-DBCP-fragment-tp3700373p3700373.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Problems with Karaf 2.2.5 integration tests
The point is that in karaf 2.x, the command is named osgi:list and not bundle:list which is only for 3.x On Fri, Feb 24, 2012 at 12:51, lennart.jore...@teliasonera.com wrote: Hello all, I have been trying for some time to get integration tests running properly in Karaf 2.2.5. The test rig is copied from what little integration tests could be found on the net. My problem is that I always receive an exception claiming that the executed command does not exist. Is this a known problem? Can anyone point me to a full Karaf integration test for Karaf 2.2.5 that actually works without relying on Karaf 3.0.0 snapshot dependencies? 2012-02-24 12:46:16,133 | WARN | rint Extender: 2 | KarArtifactInstaller | eployer.kar.KarArtifactInstaller 71 | 15 - org.apache.karaf.deployer.kar - 2.2.5 | Unable to create directory for Karaf Archive timestamps. Results may vary... __ __ / //_/ __ _/ __/ / , / __ `/ ___/ __ `/ /_ / /| |/ /_/ / / / /_/ / __/ /_/ |_|\__,_/_/ \__,_/_/ Apache Karaf (2.2.5) Hit 'tab' for a list of available commands and '[cmd] --help' for help on a specific command. Hit 'ctrl-d' or 'osgi:shutdown' to shutdown Karaf. karaf@root 2012-02-24 12:46:16,461 | WARN | rint Extender: 2 | FeaturesServiceImpl | res.internal.FeaturesServiceImpl 214 | 16 - org.apache.karaf.features.core - 2.2.5 | Feature repository doesn't have a name. The name will be mandatory in the next Karaf version. 2012-02-24 12:46:16,486 | WARN | rint Extender: 2 | FeaturesServiceImpl | res.internal.FeaturesServiceImpl 214 | 16 - org.apache.karaf.features.core - 2.2.5 | Feature repository doesn't have a name. The name will be mandatory in the next Karaf version. 2012-02-24 12:46:16,559 | WARN | rint Extender: 2 | FeaturesServiceImpl | res.internal.FeaturesServiceImpl 214 | 16 - org.apache.karaf.features.core - 2.2.5 | Feature repository doesn't have a name. The name will be mandatory in the next Karaf version. Running command: bundle:list --help org.apache.felix.gogo.runtime.CommandNotFoundException: Command not found: bundle:list at org.apache.felix.gogo.runtime.Closure.executeCmd(Closure.java:471) at org.apache.felix.gogo.runtime.Closure.executeStatement(Closure.java:400) at org.apache.felix.gogo.runtime.Pipe.run(Pipe.java:108) at org.apache.felix.gogo.runtime.Closure.execute(Closure.java:183) at org.apache.felix.gogo.runtime.Closure.execute(Closure.java:120) at org.apache.felix.gogo.runtime.CommandSessionImpl.execute(CommandSessionImpl.java:89) -- // Bästa hälsningar, // [sw. Best regards,] // // Lennart Jörelid, Systems Architect // email: lennart.jore...@teliasonera.com // cell: +46 708 507 603 // skype: jgurueurope -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: karaf/osgi start dependencies
No, but depending on the configuration it may expose a JMSCOnnectionFactory which can be obtained from the OSGi registry. On Wed, Apr 25, 2012 at 17:30, Jean-Baptiste Onofré j...@nanthrax.net wrote: Agree Achim, however, I'm not sure that ActiveMQ register a service per queue or topic for instance (only for the broker). Regards JB On 04/25/2012 05:18 PM, Achim Nierbeck wrote: Never the less you shouldn't rely on the startlevels, I suggest that your application waits on services provided by the ActiveMQ broker. This is far more safe and more OSGi like ;) regards, Achim 2012/4/25 Jean-Baptiste Onofréj...@nanthrax.net: Hi Jason, Using a feature, you can define the start-level of the bundles in the feature. Regards JB On 04/25/2012 05:12 PM, Jason wrote: Hi all, I have an application that uses Apache ActiveMQ in Karaf. I have a broker project and multiple other projects that should depend on the broker service to be started. How do I specify in the maven pom (maven-bundle-plugin) that the child projects should wait until the broker is started? Thanks, Jason -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Jean-Baptiste Onofré jbono...@apache.org http://blog.nanthrax.net Talend - http://www.talend.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: log4j
Custom appenders can be deployed by using fragments attached to the
pax-logging-service bundle fwiw.
On Thu, May 3, 2012 at 1:21 PM, Achim Nierbeck bcanh...@googlemail.comwrote:
I'm not sure what you are trying to do, but log4j as slf4j and a
couple more logging framworks are supported by Karaf.
We use Pax Logging for this, log4j is even the underlying
implementation for logging.
If you want to use your own appenders you need to take special care for
this.
Regards, Achim
2012/5/3 maaruks maris.orbid...@gmail.com:
Is it possible to use log4j in karaf ?
I have log4j classes in my bundle but I cant access them:
try {
Class? aClass = Class.forName(org.apache.log4j.Appender);
} catch (ClassNotFoundException e) {
throw new RuntimeException(e);
}
java.lang.RuntimeException: java.lang.ClassNotFoundException:
org.apache.log4j.Appender not found by
org.ops4j.pax.logging.pax-logging-api
[4]
...
Caused by: java.lang.ClassNotFoundException: org.apache.log4j.Appender
not
found by org.ops4j.pax.logging.pax-logging-api [4]
at
org.apache.felix.framework.ModuleImpl.findClassOrResourceByDelegation(ModuleImpl.java:787)
at
org.apache.felix.framework.ModuleImpl.access$400(ModuleImpl.java:71)
at
org.apache.felix.framework.ModuleImpl$ModuleClassLoader.loadClass(ModuleImpl.java:1768)
at java.lang.ClassLoader.loadClass(ClassLoader.java:247)[:1.6.0_26]
at
org.apache.felix.framework.ModuleImpl.getClassByDelegation(ModuleImpl.java:645)
at
org.apache.felix.framework.resolver.WireImpl.getClass(WireImpl.java:99)
at
org.apache.felix.framework.ModuleImpl.searchImports(ModuleImpl.java:1390)
at
org.apache.felix.framework.ModuleImpl.findClassOrResourceByDelegation(ModuleImpl.java:722)
at
org.apache.felix.framework.ModuleImpl.access$400(ModuleImpl.java:71)
at
org.apache.felix.framework.ModuleImpl$ModuleClassLoader.loadClass(ModuleImpl.java:1768)
at java.lang.ClassLoader.loadClass(ClassLoader.java:247)[:1.6.0_26]
at java.lang.Class.forName0(Native Method)[:1.6.0_26]
at java.lang.Class.forName(Class.java:169)[:1.6.0_26]
--
View this message in context:
http://karaf.922171.n3.nabble.com/log4j-tp3958839.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Apache Karaf http://karaf.apache.org/ Committer PMC
OPS4J Pax Web http://wiki.ops4j.org/display/paxweb/Pax+Web/
Committer Project Lead
OPS4J Pax Vaadin http://team.ops4j.org/wiki/display/PAXVAADIN/Home
Commiter Project Lead
blog http://notizblog.nierbeck.de/
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
Re: Is there a way to query osgi framework status (e.g. started)?
There's no such thing as a 'started' state. The osgi framework is fully
asynchronous and things can even be done on behalf of other bundles. The
only real way to know if 'something' is started is to have this thing
register an osgi service when started and wait for this service to be
registered. A lot of bundles do behave in this way so that you have
inter-bundles dependencies using services. But at the end, 'started'
depends on what meaning you put behind this word, and that mostly depends
on what you deploy.
On Thu, May 3, 2012 at 5:28 PM, bobshort jer...@check-it.ca wrote:
We have bundles that do some intensive processing. Right now they are
starting as the framework is starting and slowing startup down
considerably.
The osgi container is running on a plug computer so resources are minimal.
I want to start my processing only after the osgi container is fully
started. I've implemented a framework listener to detect when the framework
is started and then trigger my startup logic:
/public class EventListener implements FrameworkListener {
@Override
public void frameworkEvent(FrameworkEvent event) {
if (event.getType() == FrameworkEvent.STARTED) {
// Do startup logic here.
}
}
}/
This works if my bundles are installed before the framework is started, but
it obviously does not work for modules installed after the framework is
started.
Is there any way I can query the framework status from my bundle so I can
detect if the container is already fully started when the bundle is
installed? I'd like to do something like:
/
public void onBundleStarted() {
if (**test if framework already running**) {
// Do startup logic here.
}
}
/
Is this possible?
--
View this message in context:
http://karaf.922171.n3.nabble.com/Is-there-a-way-to-query-osgi-framework-status-e-g-started-tp3959588.html
Sent from the Karaf - User mailing list archive at Nabble.com.
--
Guillaume Nodet
Blog: http://gnodet.blogspot.com/
FuseSource, Integration everywhere
http://fusesource.com
Re: log4j
That's really the best thing to do in karaf, as the log configuration is handled in an osgi way, so no code should try to directly configure the logging framework. On Fri, May 4, 2012 at 1:01 PM, maaruks maris.orbid...@gmail.com wrote: Achim Nierbeck wrote I'm not sure what you are trying to do, but log4j as slf4j and a couple more logging framworks are supported by Karaf. I am migrating some code to karaf.It turns out Pax Logging doesn't allow to access all log4j classes. I removed code that is trying to access Appender. Problem solved. -- View this message in context: http://karaf.922171.n3.nabble.com/log4j-tp3958839p3961902.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Tomcat on karaf
Jetty has a long history of being lightweight and easily embeddable, with a good and responsive community, that's why it was chosen to be used in pax-web, and pax-web was really the only available solution at some point hence we have chosen it for Karaf. I don't really see any major problems deploying gemini web in Karaf though I haven't actually tested it. The question is why do you want to use Tomcat instead of Jetty ? Jetty is very mature too, though a bit less known. Doing the WAB support in OSGi is quite a lot of work and we have people here maintaining pax-web, so I don't really see the point in recreating a new container on top of Tomcat. The most important thing about Tomcat imho is that the container and management is well known, but if we'd start embedding it, you'd loose all of that anyway. On Tue, May 22, 2012 at 10:11 PM, Romain Gilles romain.gil...@gmail.com wrote: Hi all, I would like to know if any of you have run or run a tomcat web container as a Http Service and Web Application Service... Do I have to install eclipse gemini web on karaf to test it? I have had a look to pax web and the only spi provider is Jetty. Is there a reason that? Tomcat is an Apache project and the current osgi integration is done by the eclipse community and Jetty is now a eclipse project and is currently the default web container of karaf (an apache project ...)? Romain. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Bundle is not compatible with this blueprint extender
Such a problem is not really a problem ;-) One possible reason is that you have multiple blueprint extenders deployed (so multiple blueprint specs exported). An extender will only manage the blueprint bundles that are compatible (i.e. it imports the exact same package) as the extender. The main reason is to allow multiple blueprint implementations to co-exist. If you want to force the use of a given blueprint, one way is to make sure your blueprint bundle import a package which is specific to your extender, such as org.apache.aries.blueprint package in addition to the org.osgi.service.blueprint which should always be imported by blueprint bundles. The osgi framework will then be forced to wire against the aries api, making sure the bundle will be extended only by aries On Tue, Jun 19, 2012 at 6:24 PM, maaruks maris.orbid...@gmail.com wrote: I have a bundle that uses blueprint. When I try to activate it BlueprintExtender prints this: Bundle ... is not compatible with this blueprint extender Why is my bundle not compatible ? Because it uses spring 3 ? -- View this message in context: http://karaf.922171.n3.nabble.com/Bundle-is-not-compatible-with-this-blueprint-extender-tp4024898.html Sent from the Karaf - User mailing list archive at Nabble.com. -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Spring-DM web and pax web extender bundle processing order in 2.2.7
That's a quite big problem and I'm not really aware of any way to control the extender execution order. The reason is that each extender can work synchronously or asynchronously and there's no coordination between them, there's not even a way to specify such a thing. I can only see two ways out: * enhance the extenders to better cooperate, but this would require enhancing all the extenders and defining some metadata to control such an order (given anyone can write an extender, this would be better if we had a spec for that) * do the synchronization on your own, i.e. you can detect using a spring bean when the spring app will be processed and from a servlet when the web app stuff will be kicked, then make sure one is waiting for the other. The last option is the easiest one to achieve obviously, but will be really tied to the extenders your using and how they behave, as if both are started synchronously, you won't really have any way to do some synchronization here, but iirc that's not the case here. On Mon, Jun 25, 2012 at 6:21 AM, Raman Gupta rocketra...@gmail.com wrote: I just upgraded to Karaf 2.2.7 from Karaf 2.2.4 and noticed that now the Spring DM web extender and pax web extender's no longer run in the correct order. The Spring-DM extender needs to process the bundle *before* the PaxWeb extender, since until the app context is created by Spring-DM it is a non-functional web bundle. If Pax Web attempts to process it first, there is an error about the Spring context not existing when the servlet tries to initialize. A manual refresh of the bundle is required to fix the problem. One difference I can see between 2.2.4 and 2.2.7 is that the start level of the pax bundles is different. In Karaf 2.2.4, the pax bundles started at the default start level which was 60, but in 2.2.7 the war feature specifies they start at the same level as the Spring-DM bundles, which is 30. I don't know if that is the underlying problem though. Is there a way to control the order of the extender execution? If not, what is the best work-around? Regards, Raman -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Installing Hadoop bundle in Apache Felix Gogo shell
Deploying hadoop in OSGi is not a simple task, and wether you use Gogo or not isn't really the problem. Gogo just enables you to install bundles using a command instead of using the osgi api, but as I said, the problem isn't really here. So Hadoop isn't OSGi ready at all. I worked on it a few weeks ago, so you can try to build Fuse Fabric locally https://github.com/fusesource/fuse/tree/master/fabric/fabric-hadoop and try to deploy the hadoop feature (or all the bundles listed manually) https://github.com/fusesource/fuse/blob/master/fabric/fuse-fabric/src/main/resources/fabric-features.xml#L368 On Mon, Jun 25, 2012 at 10:25 AM, somya singhal 28somyasing...@gmail.com wrote: Hello I have recently learned to successfully deploy bundles with the help of Apache Felix Gogo command shell.Now i am trying to install a Hadoop bundle with the help of Apache Felix Gogo command shell.I have tried searching web pages to get clear steps for installing the same but i am unable to find.Can anyone please help me to know the correct steps for installing the Hadoop bundle with the help of Apache Felix Gogo command shell? With Regards Somya Singhal -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Spring-DM web and pax web extender bundle processing order in 2.2.7
I think Peter's solution is a nice one, as spring (and blueprint) has the ability to wait until all services are satisfied before actually creating the full spring app. That would only work if you need it that way though (i.e. spring app started after the web app). On Mon, Jun 25, 2012 at 4:14 PM, Raman Gupta rocketra...@gmail.com wrote: Thanks Guillaume and Achim... I will investigate the manual synchronization route. Does Pax-web initialize each servlet synchronously? If so, I can create a dummy servlet that initializes before the main servlet that can wait for the initialization of the Spring context. I also found this post by Peter Krien's responding to a similar query from someone on the Felix mailing list, and suggesting a similar solution: http://mail-archives.apache.org/mod_mbox/felix-users/201004.mbox/%3cf5b8b3e9-5602-4caf-9d4f-904fb947b...@aqute.biz%3E In Krien's solution, rather than manually detecting when a Spring bean has been initialized, I believe he his suggesting exposes the servlet itself as a service from the Spring context. When this service is available, the dummy servlet being initialized by the web extender sees it and proceeds (I guess this would depend on Pax-Web processing the dummy servlet first and synchronously). If that approach works, that would be some functionality that Pax-Web could use to manage this issue with requiring changes in other extenders. A manifest header could control whether Pax-Web waits for some service to be available before proceeding with the servlet initialization. This would skip the need to manually create the dummy servlet. Regards, Raman On 06/25/2012 04:12 AM, Achim Nierbeck wrote: Hi I just can second Guillaume here, there is only one more thing that crosses my mind. Pax Web needs to support Injecting OSGi Services into the Servlets, see also [1]. But as usual this kind of stuff is needed much earlier then implemented :) regards, [1] - http://team.ops4j.org/browse/PAXWEB-367 2012/6/25 Guillaume Nodet gno...@gmail.com: That's a quite big problem and I'm not really aware of any way to control the extender execution order. The reason is that each extender can work synchronously or asynchronously and there's no coordination between them, there's not even a way to specify such a thing. I can only see two ways out: * enhance the extenders to better cooperate, but this would require enhancing all the extenders and defining some metadata to control such an order (given anyone can write an extender, this would be better if we had a spec for that) * do the synchronization on your own, i.e. you can detect using a spring bean when the spring app will be processed and from a servlet when the web app stuff will be kicked, then make sure one is waiting for the other. The last option is the easiest one to achieve obviously, but will be really tied to the extenders your using and how they behave, as if both are started synchronously, you won't really have any way to do some synchronization here, but iirc that's not the case here. On Mon, Jun 25, 2012 at 6:21 AM, Raman Gupta rocketra...@gmail.com wrote: I just upgraded to Karaf 2.2.7 from Karaf 2.2.4 and noticed that now the Spring DM web extender and pax web extender's no longer run in the correct order. The Spring-DM extender needs to process the bundle *before* the PaxWeb extender, since until the app context is created by Spring-DM it is a non-functional web bundle. If Pax Web attempts to process it first, there is an error about the Spring context not existing when the servlet tries to initialize. A manual refresh of the bundle is required to fix the problem. One difference I can see between 2.2.4 and 2.2.7 is that the start level of the pax bundles is different. In Karaf 2.2.4, the pax bundles started at the default start level which was 60, but in 2.2.7 the war feature specifies they start at the same level as the Spring-DM bundles, which is 30. I don't know if that is the underlying problem though. Is there a way to control the order of the extender execution? If not, what is the best work-around? Regards, Raman -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Eventadmin as deploy seems to restart framework
That's because eventadmin is an optional dependency of a lot of bundles, including aries blueprint. FileInstall does a refresh on the bundles and it causes the blueprint extender to be re-wired, causing all blueprint applications to be restarted. If that's not done, events won't be sent by blueprint and other bundles. On Mon, Jun 25, 2012 at 7:29 PM, John Hawksley john_hawks...@intergral.com wrote: Hi folks, I seem to have a strange issue with Karaf 2.2.7. I know eventadmin can be installed as a feature (or as a boot feature), but if I deploy org.apache.felix.eventadmin-1.2.14.jar into the deploy folder, I get a strange restart behaviour. Karaf seems to: - Come up to the point I get a console prompt - Immediately stop all bundles - Immediately restart all bundles, and I get a new console banner and prompt. The log at level INFO seems to show, for instance Service MBeans being registered, then deregistered, then registered again. This may be a known restriction or caveat, but I couldn't find any other cases of this even after googling. I'm using a freshly-untarred Karaf 2.2.7 on OSX with Java 7 (1.7.0_04), and the only thing in the deploy directory is eventadmin, (Felix EA 1.2.14). If anyone can shed any light on this I'd be grateful; we're trying to come up with a bundle set which doesn't rely on features, so we'd like to deploy eventadmin too. If this seems to be unknown, naturally I'll open a bug for it. Many thanks everyone, -John -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Spring-DM web and pax web extender bundle processing order in 2.2.7
If you use a spring web app, do you really need to use spring-dm in addition ? Why don't you just use the spring app ? On Tue, Jun 26, 2012 at 1:23 AM, Raman Gupta rocketra...@gmail.com wrote: On 06/25/2012 03:21 AM, Guillaume Nodet wrote: That's a quite big problem and I'm not really aware of any way to control the extender execution order. The reason is that each extender can work synchronously or asynchronously and there's no coordination between them, there's not even a way to specify such a thing. I can only see two ways out: * enhance the extenders to better cooperate, but this would require enhancing all the extenders and defining some metadata to control such an order (given anyone can write an extender, this would be better if we had a spec for that) * do the synchronization on your own, i.e. you can detect using a spring bean when the spring app will be processed and from a servlet when the web app stuff will be kicked, then make sure one is waiting for the other. The last option is the easiest one to achieve obviously, but will be really tied to the extenders your using and how they behave, as if both are started synchronously, you won't really have any way to do some synchronization here, but iirc that's not the case here. I'm trying to do the second option (manual synchronization) as described here. It isn't working as Pax Web itself attempts to initialize the Spring context due to the Spring ContextLoaderListener that is defined in web.xml, which then breaks the manual synchronization. Here is the stack for the Pax Web initialization of the context: http://pastebin.com/raw.php?i=RTunRTkF From the stack, the general path to the bean being initialized is: swissbox BundleWatcher - WebXmlObserver - WebAppPublisher - Pax Web Jetty - ContextLoaderListener - ExtenderOrderingBean.init() Then when the servlet is loaded by Pax Web, ExtenderOrderingBean thinks the context is good since it was created by Pax Web, and the stack which started this whole mess is again produced: http://pastebin.com/raw.php?i=PJvMqWRV I actually don't see where the Spring-DM extender loads the bean at all... So still looking for solutions... IMO, the behavior of Karaf 2.2.4 was perfect in this situation, and required no such hackish work-arounds. Regards, Raman -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Spring-DM web and pax web extender bundle processing order in 2.2.7
If you want to use spring-dm and spring-mvc, you want to look at http://static.springsource.org/osgi/docs/1.1.0-m2/reference/html/web.html There's a special application context class designed for this it seems: org.springframework.osgi.web.context.support.OsgiBundleXmlWebApplicationContext On Tue, Jun 26, 2012 at 4:42 PM, Raman Gupta rocketra...@gmail.com wrote: On 06/26/2012 07:17 AM, Guillaume Nodet wrote: If you use a spring web app, do you really need to use spring-dm in addition ? Why don't you just use the spring app ? I'm not sure what you mean... Spring-DM is what ties in the OSGi microservices into the web bundle's Spring context, right? How am I supposed to do that without using Spring-DM or Gemini Blueprint, or doing it manually using the OSGi api's? On Tue, Jun 26, 2012 at 1:23 AM, Raman Gupta rocketra...@gmail.com wrote: On 06/25/2012 03:21 AM, Guillaume Nodet wrote: That's a quite big problem and I'm not really aware of any way to control the extender execution order. The reason is that each extender can work synchronously or asynchronously and there's no coordination between them, there's not even a way to specify such a thing. I can only see two ways out: * enhance the extenders to better cooperate, but this would require enhancing all the extenders and defining some metadata to control such an order (given anyone can write an extender, this would be better if we had a spec for that) * do the synchronization on your own, i.e. you can detect using a spring bean when the spring app will be processed and from a servlet when the web app stuff will be kicked, then make sure one is waiting for the other. The last option is the easiest one to achieve obviously, but will be really tied to the extenders your using and how they behave, as if both are started synchronously, you won't really have any way to do some synchronization here, but iirc that's not the case here. I'm trying to do the second option (manual synchronization) as described here. It isn't working as Pax Web itself attempts to initialize the Spring context due to the Spring ContextLoaderListener that is defined in web.xml, which then breaks the manual synchronization. Here is the stack for the Pax Web initialization of the context: http://pastebin.com/raw.php?i=RTunRTkF From the stack, the general path to the bean being initialized is: swissbox BundleWatcher - WebXmlObserver - WebAppPublisher - Pax Web Jetty - ContextLoaderListener - ExtenderOrderingBean.init() Then when the servlet is loaded by Pax Web, ExtenderOrderingBean thinks the context is good since it was created by Pax Web, and the stack which started this whole mess is again produced: http://pastebin.com/raw.php?i=PJvMqWRV I actually don't see where the Spring-DM extender loads the bean at all... So still looking for solutions... IMO, the behavior of Karaf 2.2.4 was perfect in this situation, and required no such hackish work-arounds. Regards, Raman -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
Re: Bundles for deployment of hadoop service in osgi container
Have you tried the links I gave you ? On Wed, Jun 27, 2012 at 10:58 AM, somya singhal 28somyasing...@gmail.com wrote: Hello I have recently coma across a link -: http://search-hadoop.com/m/7TZE59pm6vsubj=Re+PROPOSAL+Hadoop+OSGi+compliant+and+Apache+Karaf+features I am trying to install hadoop bundle in osgi for quite a few days.But i am unable to do so.Anyone can please suggest me as it is written in the link,from where can i get Hadoop modules (common, annotations, hdfs, mapreduce, etc)??? Somya Singhal Btech(4th year,csi) IIT ROORKEE -- Guillaume Nodet Blog: http://gnodet.blogspot.com/ FuseSource, Integration everywhere http://fusesource.com
