[jira] [Commented] (KARAF-4260) Setting karaf.clean.all = true breaks service wrapper service script
[ https://issues.apache.org/jira/browse/KARAF-4260?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15434315#comment-15434315 ] Ralf Steppacher commented on KARAF-4260: We put the PID file into {{KARAF_BASE}} and had no issues with that location. > Setting karaf.clean.all = true breaks service wrapper service script > > > Key: KARAF-4260 > URL: https://issues.apache.org/jira/browse/KARAF-4260 > Project: Karaf > Issue Type: Bug > Components: karaf-config >Affects Versions: 4.0.3 >Reporter: Ralf Steppacher >Assignee: Jean-Baptiste Onofré >Priority: Minor > Fix For: 4.1.0, 4.0.7 > > > The Karaf service wrapper script is generated such that the PID file is > created in {{$KARAF_HOME/data}}. When setting {{karaf.clean.all = true}} then > the PID file created by the service script gets deleted together with the > data directory. As a result of this the service script reports Karaf as not > running and it is not possible to stop the process via the script. > The PID file location probably should be outside the data directory by > default. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (KARAF-4261) Bundle start-level seems to be ignored at Karaf restart
[ https://issues.apache.org/jira/browse/KARAF-4261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15405449#comment-15405449 ] Ralf Steppacher commented on KARAF-4261: [~ch...@die-schneider.net], thanks for your comment and apologies for my long response times. Yes, as far as I can tell the problem occurs only when deploying a KAR file via the {{deploy}} folder. Not being able to use a KAR with the deploy folder will make the instructions for our operations team for deployment and upgrades that much more elaborate and error prone. It would be really great of deploying and undeploying a KAR via the deploy folder could be fixed. > Bundle start-level seems to be ignored at Karaf restart > --- > > Key: KARAF-4261 > URL: https://issues.apache.org/jira/browse/KARAF-4261 > Project: Karaf > Issue Type: Bug > Components: karaf-core >Affects Versions: 4.0.3 >Reporter: Ralf Steppacher > Fix For: 4.1.0, 4.0.6 > > > AS a workaround for CAMEL-9483 I have set a start-level for my bundles > deployed as part of my features. This works as expected during initial > deployment (order of deployment is according to the start levels I set), but > not during sub-sequent starts of Karaf. It appears the start-level of the > bundles is ignored, meaning the order of deployment of my bundles is more or > less random and I observe the issues described in CAMEL-9483 again. > {{bundle:list}} shows my bundles with the start-levels they have been > originally deployed with, though. > As a workaround I set {{karaf.clean.all = true}} in system.properties. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Comment Edited] (KARAF-4261) Bundle start-level seems to be ignored at Karaf restart
[ https://issues.apache.org/jira/browse/KARAF-4261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15352446#comment-15352446 ] Ralf Steppacher edited comment on KARAF-4261 at 6/28/16 6:28 AM: - I am using Felix. was (Author: ralfsteppacher): I am using Felix. > Bundle start-level seems to be ignored at Karaf restart > --- > > Key: KARAF-4261 > URL: https://issues.apache.org/jira/browse/KARAF-4261 > Project: Karaf > Issue Type: Bug > Components: karaf-core >Affects Versions: 4.0.3 >Reporter: Ralf Steppacher > Fix For: 4.1.0, 4.0.6 > > > AS a workaround for CAMEL-9483 I have set a start-level for my bundles > deployed as part of my features. This works as expected during initial > deployment (order of deployment is according to the start levels I set), but > not during sub-sequent starts of Karaf. It appears the start-level of the > bundles is ignored, meaning the order of deployment of my bundles is more or > less random and I observe the issues described in CAMEL-9483 again. > {{bundle:list}} shows my bundles with the start-levels they have been > originally deployed with, though. > As a workaround I set {{karaf.clean.all = true}} in system.properties. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (KARAF-4261) Bundle start-level seems to be ignored at Karaf restart
[ https://issues.apache.org/jira/browse/KARAF-4261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15352446#comment-15352446 ] Ralf Steppacher commented on KARAF-4261: I am using Felix. > Bundle start-level seems to be ignored at Karaf restart > --- > > Key: KARAF-4261 > URL: https://issues.apache.org/jira/browse/KARAF-4261 > Project: Karaf > Issue Type: Bug > Components: karaf-core >Affects Versions: 4.0.3 >Reporter: Ralf Steppacher > Fix For: 4.1.0, 4.0.6 > > > AS a workaround for CAMEL-9483 I have set a start-level for my bundles > deployed as part of my features. This works as expected during initial > deployment (order of deployment is according to the start levels I set), but > not during sub-sequent starts of Karaf. It appears the start-level of the > bundles is ignored, meaning the order of deployment of my bundles is more or > less random and I observe the issues described in CAMEL-9483 again. > {{bundle:list}} shows my bundles with the start-levels they have been > originally deployed with, though. > As a workaround I set {{karaf.clean.all = true}} in system.properties. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Updated] (KARAF-4522) Instantiating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn
[ https://issues.apache.org/jira/browse/KARAF-4522?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ralf Steppacher updated KARAF-4522: --- Description: I am trying to deploy a bundle that starts a Camel context via a Blueprint context file. The bundle depends on a custom Camel component to subscribe to LDAP updates, which again depends on the UnboundID SDK. {noformat} bundle: camel context --> bundle: custom camel component --> bundle: UnboundID SDK {noformat} The UnboundID SDK offers a call-back interface {{DisconnectHandler}} which I implement in the custom camel component. In the call-back implementation I spawn a new thread that attempts to re-establish the connection at a certain interval. Instantiating this thread fails consistently on the first attempt, but always succeeds on the second. If I wrap the call to {{new ReconnectionThread}} in a loop like so: {code:java} success = false; while(!success) { try { ReconnectThread rt = new ReconnectThread(...); success = true; } catch (Exception e) { LOG.warn("Failed to spawn LDAP re-connection thread.", e); } } {code} Then the first attempt fails consistently while the second time around it consistently succeeds. The exception I receive is the following: {noformat} 2016-05-09 08:37:24,492 | WARN | dap:389 (closed) | ReconnectingDisconnectHandler| 53 - ch.vivates.ams.camel-uldap - 4.0.9.SNAPSHOT | | Failed to spawn LDAP re-connection thread. java.lang.IllegalStateException: Unknown protocol: mvn at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:482)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:474)[org.apache.felix.framework-5.4.0.jar:] at java.net.URL.toExternalForm(URL.java:929)[:1.8.0_77] at java.net.URL.toString(URL.java:915)[:1.8.0_77] at java.lang.ClassLoader.defineClassSourceLocation(ClassLoader.java:678)[:1.8.0_77] at java.lang.ClassLoader.defineClass(ClassLoader.java:762)[:1.8.0_77] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.defineClass(BundleWiringImpl.java:2370)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.findClass(BundleWiringImpl.java:2154)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.findClassOrResourceByDelegation(BundleWiringImpl.java:1542)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.access$400(BundleWiringImpl.java:79)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.loadClass(BundleWiringImpl.java:2018)[org.apache.felix.framework-5.4.0.jar:] at java.lang.ClassLoader.loadClass(ClassLoader.java:357)[:1.8.0_77] at ch.vivates.ams.camel.compoment.uldap.unboundid.ReconnectingDisconnectHandler.handleDisconnect(ReconnectingDisconnectHandler.java:44)[53:ch.vivates.ams.camel-uldap:4.0.9.SNAPSHOT] at com.unboundid.ldap.sdk.DisconnectInfo.notifyDisconnectHandler(DisconnectInfo.java:149)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionInternals.close(LDAPConnectionInternals.java:665)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnection.setClosed(LDAPConnection.java:4431)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.closeInternal(LDAPConnectionReader.java:1089)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.run(LDAPConnectionReader.java:444)[56:com.unboundid.ldap.sdk.se:3.1.1] {noformat} It appears the problem is limited to bundle deployment time. If the LDAP server is available while the bundle gets deployed and only fails later, then the reconnection thread is spawned on the first attempt. Another thing I have noticed: The creation of the reconnection thread only fails when I attempt to spawn it from the disconnect handler, which is driven by a thread associated to the LDAP connection ({{LDAPConnectionReader}}). If I try to spawn my reconnect-thread from the thread that starts the camel component, then this works on the first attempt as well. The feature file I am using to deploy the bundles is simple: {code:xml} http://karaf.apache.org/xmlns/features/v1.3.0; name="Features"> Camel endpoint to subscribe with an OpenLDAP server and convert updates to the LDAP into Camel exchanges. camel-core mvn:ch.vivates.ams/camel-uldap/4.0.9-SNAPSHOT mvn:com.unboundid/unboundid-ldapsdk/3.1.1 Replicates the HPD LDAP into the graph DB transaction camel-core camel-blueprint base titan-client
[jira] [Updated] (KARAF-4522) Instantiating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn
[ https://issues.apache.org/jira/browse/KARAF-4522?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ralf Steppacher updated KARAF-4522: --- Description: I am trying to deploy a bundle that starts a Camel context via a Blueprint context file. The bundle depends on a custom Camel component to subscribe to LDAP updates, which again depends on the UnboundID SDK. {noformat} bundle: camel context --> bundle: custom camel component --> bundle: UnboundID SDK {noformat} The UnboundID SDK offers a call-back interface {{DisconnectHandler}} which I implement in the custom camel component. In the call-back implementation I spawn a new thread that attempts to re-establish the connection at a certain interval. Instantiating this thread fails consistently on the first attempt, but always succeeds on the second. If I wrap the call to {{new ReconnectionThread}} in a loop like so: {code:java} success = false; while(!success) { try { ReconnectThread rt = new ReconnectThread(...); success = true; } catch (Exception e) { LOG.warn("Failed to spawn LDAP re-connection thread.", e); } } {code} Then the first attempt fails consistently while the second time around it consistently succeeds. The exception I receive is the following: {noformat} 2016-05-09 08:37:24,492 | WARN | dap:389 (closed) | ReconnectingDisconnectHandler| 53 - ch.vivates.ams.camel-uldap - 4.0.9.SNAPSHOT | | Failed to spawn LDAP re-connection thread. java.lang.IllegalStateException: Unknown protocol: mvn at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:482)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:474)[org.apache.felix.framework-5.4.0.jar:] at java.net.URL.toExternalForm(URL.java:929)[:1.8.0_77] at java.net.URL.toString(URL.java:915)[:1.8.0_77] at java.lang.ClassLoader.defineClassSourceLocation(ClassLoader.java:678)[:1.8.0_77] at java.lang.ClassLoader.defineClass(ClassLoader.java:762)[:1.8.0_77] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.defineClass(BundleWiringImpl.java:2370)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.findClass(BundleWiringImpl.java:2154)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.findClassOrResourceByDelegation(BundleWiringImpl.java:1542)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.access$400(BundleWiringImpl.java:79)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.loadClass(BundleWiringImpl.java:2018)[org.apache.felix.framework-5.4.0.jar:] at java.lang.ClassLoader.loadClass(ClassLoader.java:357)[:1.8.0_77] at ch.vivates.ams.camel.compoment.uldap.unboundid.ReconnectingDisconnectHandler.handleDisconnect(ReconnectingDisconnectHandler.java:44)[53:ch.vivates.ams.camel-uldap:4.0.9.SNAPSHOT] at com.unboundid.ldap.sdk.DisconnectInfo.notifyDisconnectHandler(DisconnectInfo.java:149)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionInternals.close(LDAPConnectionInternals.java:665)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnection.setClosed(LDAPConnection.java:4431)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.closeInternal(LDAPConnectionReader.java:1089)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.run(LDAPConnectionReader.java:444)[56:com.unboundid.ldap.sdk.se:3.1.1] {noformat} It appears the problem is limited to bundle deployment time. If the LDAP server is available while the bundle gets deployed and only fails later, then the reconnection thread is spawned on the first attempt. Another thing I have noticed: The creation of the reconnection thread only fails when I attempt to spawn it from the disconnect handler, which is driven by a thread associated to the LDAP connection ({{LDAPConnectionReader}}). If I try to spawn my reconnect-thread from the thread that starts the camel component, then this works on the first attempt as well. The feature file I am using to deploy the bundles is simple: {code:xml} http://karaf.apache.org/xmlns/features/v1.3.0; name="Features"> Camel endpoint to subscribe with an OpenLDAP server and convert updates to the LDAP into Camel exchanges. camel-core mvn:ch.vivates.ams/camel-uldap/4.0.9-SNAPSHOT mvn:com.unboundid/unboundid-ldapsdk/3.1.1 {code} was: I am trying to deploy a bundle that starts a Camel context via a Blueprint context file. The bundle depends on a custom Camel component
[jira] [Updated] (KARAF-4522) Instantiating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn
[ https://issues.apache.org/jira/browse/KARAF-4522?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ralf Steppacher updated KARAF-4522: --- Description: I am trying to deploy a bundle that starts a Camel context via a Blueprint context file. The bundle depends on a custom Camel component to subscribe to LDAP updates, which again depends on the UnboundID SDK. {noformat} bundle: camel context --> bundle: custom camel component --> bundle: UnboundID SDK {noformat} The UnboundID SDK offers a call-back interface {{DisconnectHandler}} which I implement in the custom camel component. In the call-back implementation I spawn a new thread that attempts to re-establish the connection at a certain interval. Instantiating this thread fails consistently on the first attempt, but always succeeds on the second. If I wrap the call to {{new ReconnectionThread}} in a loop like so: {code:java} success = false; while(!success) { try { ReconnectThread rt = new ReconnectThread(...); success = true; } catch (Exception e) { LOG.warn("Failed to spawn LDAP re-connection thread.", e); } } {code} Then the first attempt fails consistently while the second time around it consistently succeeds. The exception I receive is the following: {noformat} 2016-05-09 08:37:24,492 | WARN | dap:389 (closed) | ReconnectingDisconnectHandler| 53 - ch.vivates.ams.camel-uldap - 4.0.9.SNAPSHOT | | Failed to spawn LDAP re-connection thread. java.lang.IllegalStateException: Unknown protocol: mvn at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:482)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:474)[org.apache.felix.framework-5.4.0.jar:] at java.net.URL.toExternalForm(URL.java:929)[:1.8.0_77] at java.net.URL.toString(URL.java:915)[:1.8.0_77] at java.lang.ClassLoader.defineClassSourceLocation(ClassLoader.java:678)[:1.8.0_77] at java.lang.ClassLoader.defineClass(ClassLoader.java:762)[:1.8.0_77] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.defineClass(BundleWiringImpl.java:2370)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.findClass(BundleWiringImpl.java:2154)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.findClassOrResourceByDelegation(BundleWiringImpl.java:1542)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.access$400(BundleWiringImpl.java:79)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.loadClass(BundleWiringImpl.java:2018)[org.apache.felix.framework-5.4.0.jar:] at java.lang.ClassLoader.loadClass(ClassLoader.java:357)[:1.8.0_77] at ch.vivates.ams.camel.compoment.uldap.unboundid.ReconnectingDisconnectHandler.handleDisconnect(ReconnectingDisconnectHandler.java:44)[53:ch.vivates.ams.camel-uldap:4.0.9.SNAPSHOT] at com.unboundid.ldap.sdk.DisconnectInfo.notifyDisconnectHandler(DisconnectInfo.java:149)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionInternals.close(LDAPConnectionInternals.java:665)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnection.setClosed(LDAPConnection.java:4431)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.closeInternal(LDAPConnectionReader.java:1089)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.run(LDAPConnectionReader.java:444)[56:com.unboundid.ldap.sdk.se:3.1.1] {noformat} It appears the problem is limited to bundle deployment time. If the LDAP server is available while the bundle gets deployed and only fails later, then the reconnection thread is spawned on the first attempt. Another thing I have noticed: The creation of the reconnection thread only fails when I attempt to spawn it from the disconnect handler, which is driven by a thread associated to the LDAP connection ({{LDAPConnectionReader}}). If I try to spawn my reconnect-thread from the thread that starts the camel component, then this works on the first attempt as well. was: I am trying to deploy a bundle that starts a Camel context via a Blueprint context file. The bundle depends on a custom Camel component to subscribe to LDAP updates, which again depends on the UnboundID SDK. {noformat} bundle: camel context --> bundle: custom camel component --> bundle: UnboundID SDK {noformat} The UnboundID SDK offers a call-back interface {{DisconnectHandler}} which I implement in the custom camel component. In the call-back implementation I spawn a new thread that attempts to re-establish the connection at a
[jira] [Updated] (KARAF-4522) Instantiating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn
[ https://issues.apache.org/jira/browse/KARAF-4522?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ralf Steppacher updated KARAF-4522: --- Summary: Instantiating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn (was: Instanciating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn) > Instantiating class fails on first try but succeeds on second attempt: > java.lang.IllegalStateException: Unknown protocol: mvn > - > > Key: KARAF-4522 > URL: https://issues.apache.org/jira/browse/KARAF-4522 > Project: Karaf > Issue Type: Bug > Components: karaf-core >Affects Versions: 4.0.5 > Environment: OSX El Capitan, IBM JDK 8, Camel 2.16.3, Blueprint >Reporter: Ralf Steppacher > > I am trying to deploy a bundle that starts a Camel context via a Blueprint > context file. The bundle depends on a custom Camel component to subscribe to > LDAP updates, which again depends on the UnboundID SDK. > {noformat} > bundle: camel context --> bundle: custom camel component --> bundle: > UnboundID SDK > {noformat} > The UnboundID SDK offers a call-back interface {{DisconnectHandler}} which I > implement in the custom camel component. In the call-back implementation I > spawn a new thread that attempts to re-establish the connection at a certain > interval. Instantiating this thread fails consistently on the first attempt, > but always succeeds on the second. > If I wrap the call to {{new ReconnectionThread}} in a loop like so: > {code:java} > success = false; > while(!success) { > try { > ReconnectThread rt = new ReconnectThread(...); > success = true; > } catch (Exception e) { > LOG.warn("Failed to spawn LDAP re-connection thread.", e); > } > } > {code} > Then the first attempt fails consistently while the second time around it > consistently succeeds. > The exception I receive is the following: > {noformat} > 2016-05-09 08:37:24,492 | WARN | dap:389 (closed) | > ReconnectingDisconnectHandler| 53 - ch.vivates.ams.camel-uldap - > 4.0.9.SNAPSHOT | | Failed to spawn LDAP re-connection thread. > java.lang.IllegalStateException: Unknown protocol: mvn > at > org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:482)[org.apache.felix.framework-5.4.0.jar:] > at > org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:474)[org.apache.felix.framework-5.4.0.jar:] > at java.net.URL.toExternalForm(URL.java:929)[:1.8.0_77] > at java.net.URL.toString(URL.java:915)[:1.8.0_77] > at > java.lang.ClassLoader.defineClassSourceLocation(ClassLoader.java:678)[:1.8.0_77] > at java.lang.ClassLoader.defineClass(ClassLoader.java:762)[:1.8.0_77] > at > org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.defineClass(BundleWiringImpl.java:2370)[org.apache.felix.framework-5.4.0.jar:] > at > org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.findClass(BundleWiringImpl.java:2154)[org.apache.felix.framework-5.4.0.jar:] > at > org.apache.felix.framework.BundleWiringImpl.findClassOrResourceByDelegation(BundleWiringImpl.java:1542)[org.apache.felix.framework-5.4.0.jar:] > at > org.apache.felix.framework.BundleWiringImpl.access$400(BundleWiringImpl.java:79)[org.apache.felix.framework-5.4.0.jar:] > at > org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.loadClass(BundleWiringImpl.java:2018)[org.apache.felix.framework-5.4.0.jar:] > at java.lang.ClassLoader.loadClass(ClassLoader.java:357)[:1.8.0_77] > at > ch.vivates.ams.camel.compoment.uldap.unboundid.ReconnectingDisconnectHandler.handleDisconnect(ReconnectingDisconnectHandler.java:44)[53:ch.vivates.ams.camel-uldap:4.0.9.SNAPSHOT] > at > com.unboundid.ldap.sdk.DisconnectInfo.notifyDisconnectHandler(DisconnectInfo.java:149)[56:com.unboundid.ldap.sdk.se:3.1.1] > at > com.unboundid.ldap.sdk.LDAPConnectionInternals.close(LDAPConnectionInternals.java:665)[56:com.unboundid.ldap.sdk.se:3.1.1] > at > com.unboundid.ldap.sdk.LDAPConnection.setClosed(LDAPConnection.java:4431)[56:com.unboundid.ldap.sdk.se:3.1.1] > at > com.unboundid.ldap.sdk.LDAPConnectionReader.closeInternal(LDAPConnectionReader.java:1089)[56:com.unboundid.ldap.sdk.se:3.1.1] > at > com.unboundid.ldap.sdk.LDAPConnectionReader.run(LDAPConnectionReader.java:444)[56:com.unboundid.ldap.sdk.se:3.1.1] > {noformat} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Created] (KARAF-4522) Instanciating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn
Ralf Steppacher created KARAF-4522: -- Summary: Instanciating class fails on first try but succeeds on second attempt: java.lang.IllegalStateException: Unknown protocol: mvn Key: KARAF-4522 URL: https://issues.apache.org/jira/browse/KARAF-4522 Project: Karaf Issue Type: Bug Components: karaf-core Affects Versions: 4.0.5 Environment: OSX El Capitan, IBM JDK 8, Camel 2.16.3, Blueprint Reporter: Ralf Steppacher I am trying to deploy a bundle that starts a Camel context via a Blueprint context file. The bundle depends on a custom Camel component to subscribe to LDAP updates, which again depends on the UnboundID SDK. {noformat} bundle: camel context --> bundle: custom camel component --> bundle: UnboundID SDK {noformat} The UnboundID SDK offers a call-back interface {{DisconnectHandler}} which I implement in the custom camel component. In the call-back implementation I spawn a new thread that attempts to re-establish the connection at a certain interval. Instantiating this thread fails consistently on the first attempt, but always succeeds on the second. If I wrap the call to {{new ReconnectionThread}} in a loop like so: {code:java} success = false; while(!success) { try { ReconnectThread rt = new ReconnectThread(...); success = true; } catch (Exception e) { LOG.warn("Failed to spawn LDAP re-connection thread.", e); } } {code} Then the first attempt fails consistently while the second time around it consistently succeeds. The exception I receive is the following: {noformat} 2016-05-09 08:37:24,492 | WARN | dap:389 (closed) | ReconnectingDisconnectHandler| 53 - ch.vivates.ams.camel-uldap - 4.0.9.SNAPSHOT | | Failed to spawn LDAP re-connection thread. java.lang.IllegalStateException: Unknown protocol: mvn at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:482)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.URLHandlersStreamHandlerProxy.toExternalForm(URLHandlersStreamHandlerProxy.java:474)[org.apache.felix.framework-5.4.0.jar:] at java.net.URL.toExternalForm(URL.java:929)[:1.8.0_77] at java.net.URL.toString(URL.java:915)[:1.8.0_77] at java.lang.ClassLoader.defineClassSourceLocation(ClassLoader.java:678)[:1.8.0_77] at java.lang.ClassLoader.defineClass(ClassLoader.java:762)[:1.8.0_77] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.defineClass(BundleWiringImpl.java:2370)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.findClass(BundleWiringImpl.java:2154)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.findClassOrResourceByDelegation(BundleWiringImpl.java:1542)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl.access$400(BundleWiringImpl.java:79)[org.apache.felix.framework-5.4.0.jar:] at org.apache.felix.framework.BundleWiringImpl$BundleClassLoader.loadClass(BundleWiringImpl.java:2018)[org.apache.felix.framework-5.4.0.jar:] at java.lang.ClassLoader.loadClass(ClassLoader.java:357)[:1.8.0_77] at ch.vivates.ams.camel.compoment.uldap.unboundid.ReconnectingDisconnectHandler.handleDisconnect(ReconnectingDisconnectHandler.java:44)[53:ch.vivates.ams.camel-uldap:4.0.9.SNAPSHOT] at com.unboundid.ldap.sdk.DisconnectInfo.notifyDisconnectHandler(DisconnectInfo.java:149)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionInternals.close(LDAPConnectionInternals.java:665)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnection.setClosed(LDAPConnection.java:4431)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.closeInternal(LDAPConnectionReader.java:1089)[56:com.unboundid.ldap.sdk.se:3.1.1] at com.unboundid.ldap.sdk.LDAPConnectionReader.run(LDAPConnectionReader.java:444)[56:com.unboundid.ldap.sdk.se:3.1.1] {noformat} -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (KARAF-4261) Bundle start-level seems to be ignored at Karaf restart
[ https://issues.apache.org/jira/browse/KARAF-4261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15093550#comment-15093550 ] Ralf Steppacher commented on KARAF-4261: I tried on a second machine (3 start/stop cycles). It appears that the start/stop order is the same per Karaf instance, if that is at all possible with {{karaf.clean.all = true}}. On the second machine the data access bundle is consistently started last (3 of 3 times) but is shut as the second bundle after the _PEP_ bundle (3 of 3 times). > Bundle start-level seems to be ignored at Karaf restart > --- > > Key: KARAF-4261 > URL: https://issues.apache.org/jira/browse/KARAF-4261 > Project: Karaf > Issue Type: Bug > Components: karaf-core >Affects Versions: 4.0.3 >Reporter: Ralf Steppacher > > AS a workaround for CAMEL-9483 I have set a start-level for my bundles > deployed as part of my features. This works as expected during initial > deployment (order of deployment is according to the start levels I set), but > not during sub-sequent starts of Karaf. It appears the start-level of the > bundles is ignored, meaning the order of deployment of my bundles is more or > less random and I observe the issues described in CAMEL-9483 again. > {{bundle:list}} shows my bundles with the start-levels they have been > originally deployed with, though. > As a workaround I set {{karaf.clean.all = true}} in system.properties. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (KARAF-4261) Bundle start-level seems to be ignored at Karaf restart
[ https://issues.apache.org/jira/browse/KARAF-4261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15093503#comment-15093503 ] Ralf Steppacher commented on KARAF-4261: It looks like the bundle start-level is also not fully honored during the initial KAR deployment either. Or my understanding of the start-level is wrong. I have changed the order in which my bundles should be started such that my data access bundle starts before the webservices start and shuts down after the webservices have shut down. However, the change is not reflected in the start/stop order of the bundles. The data access bundle (_Graph Manager_, id 206) is still consistently started last and shut down first. Though looking at the start-levels it should be the other way round: {noformat} karaf@root()> bundle:list | grep 4.0.3.SNAP 52 | Active | 88 | 4.0.3.SNAPSHOT | Base 60 | Active | 89 | 4.0.3.SNAPSHOT | Patient webservice implementation 63 | Active | 89 | 4.0.3.SNAPSHOT | Camel UnboundIDLDAP Component 65 | Active | 91 | 4.0.3.SNAPSHOT | Webservice Health Professionals 152 | Active | 91 | 4.0.3.SNAPSHOT | REST API 167 | Active | 91 | 4.0.3.SNAPSHOT | Webservice Patients 168 | Active | 91 | 4.0.3.SNAPSHOT | Webservice Patient Registration 177 | Active | 91 | 4.0.3.SNAPSHOT | PEP 206 | Active | 90 | 4.0.3.SNAPSHOT | Graph Manager karaf@root()> {noformat} > Bundle start-level seems to be ignored at Karaf restart > --- > > Key: KARAF-4261 > URL: https://issues.apache.org/jira/browse/KARAF-4261 > Project: Karaf > Issue Type: Bug > Components: karaf-core >Affects Versions: 4.0.3 >Reporter: Ralf Steppacher > > AS a workaround for CAMEL-9483 I have set a start-level for my bundles > deployed as part of my features. This works as expected during initial > deployment (order of deployment is according to the start levels I set), but > not during sub-sequent starts of Karaf. It appears the start-level of the > bundles is ignored, meaning the order of deployment of my bundles is more or > less random and I observe the issues described in CAMEL-9483 again. > {{bundle:list}} shows my bundles with the start-levels they have been > originally deployed with, though. > As a workaround I set {{karaf.clean.all = true}} in system.properties. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Created] (KARAF-4261) Bundle start-level seems to be ignored at Karaf restart
Ralf Steppacher created KARAF-4261: -- Summary: Bundle start-level seems to be ignored at Karaf restart Key: KARAF-4261 URL: https://issues.apache.org/jira/browse/KARAF-4261 Project: Karaf Issue Type: Bug Components: karaf-core Affects Versions: 4.0.3 Reporter: Ralf Steppacher AS a workaround for CAMEL-9483 I have set a start-level for my bundles deployed as part of my features. This works as expected during initial deployment (order of deployment is according to the start levels I set), but not during sub-sequent starts of Karaf. It appears the start-level of the bundles is ignored, meaning the order of deployment of my bundles is more or less random and I observe the issues described in CAMEL-9483 again. {{bundle:list}} shows my bundles with the start-levels they have been originally deployed with, though. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Created] (KARAF-4260) Setting karaf.clean.all = true breaks service wrapper service script
Ralf Steppacher created KARAF-4260: -- Summary: Setting karaf.clean.all = true breaks service wrapper service script Key: KARAF-4260 URL: https://issues.apache.org/jira/browse/KARAF-4260 Project: Karaf Issue Type: Bug Components: karaf-config Affects Versions: 4.0.3 Reporter: Ralf Steppacher Priority: Minor The Karaf service wrapper script is generated such that the PID file is created in {{$KARAF_HOME/data}}. When setting {{karaf.clean.all = true}} then the PID file created by the service script gets deleted together with the data directory. As a result of this the service script reports Karaf as not running and it is not possible to stop the process via the script. The PID file location probably should be outside the data directory by default. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Updated] (KARAF-4261) Bundle start-level seems to be ignored at Karaf restart
[ https://issues.apache.org/jira/browse/KARAF-4261?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ] Ralf Steppacher updated KARAF-4261: --- Description: AS a workaround for CAMEL-9483 I have set a start-level for my bundles deployed as part of my features. This works as expected during initial deployment (order of deployment is according to the start levels I set), but not during sub-sequent starts of Karaf. It appears the start-level of the bundles is ignored, meaning the order of deployment of my bundles is more or less random and I observe the issues described in CAMEL-9483 again. {{bundle:list}} shows my bundles with the start-levels they have been originally deployed with, though. As a workaround I set {{karaf.clean.all = true}} in system.properties. was: AS a workaround for CAMEL-9483 I have set a start-level for my bundles deployed as part of my features. This works as expected during initial deployment (order of deployment is according to the start levels I set), but not during sub-sequent starts of Karaf. It appears the start-level of the bundles is ignored, meaning the order of deployment of my bundles is more or less random and I observe the issues described in CAMEL-9483 again. {{bundle:list}} shows my bundles with the start-levels they have been originally deployed with, though. > Bundle start-level seems to be ignored at Karaf restart > --- > > Key: KARAF-4261 > URL: https://issues.apache.org/jira/browse/KARAF-4261 > Project: Karaf > Issue Type: Bug > Components: karaf-core >Affects Versions: 4.0.3 >Reporter: Ralf Steppacher > > AS a workaround for CAMEL-9483 I have set a start-level for my bundles > deployed as part of my features. This works as expected during initial > deployment (order of deployment is according to the start levels I set), but > not during sub-sequent starts of Karaf. It appears the start-level of the > bundles is ignored, meaning the order of deployment of my bundles is more or > less random and I observe the issues described in CAMEL-9483 again. > {{bundle:list}} shows my bundles with the start-levels they have been > originally deployed with, though. > As a workaround I set {{karaf.clean.all = true}} in system.properties. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (KARAF-3286) karaf-maven-plugin does not merge bundle entry with URL parameters
[ https://issues.apache.org/jira/browse/KARAF-3286?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14173988#comment-14173988 ] Ralf Steppacher commented on KARAF-3286: I have the syntax from the Karaf [deployer documention|http://karaf.apache.org/manual/latest/users-guide/deployers.html] (bottom of the page). And it works if I use it like that in the feature.xml. I will try your suggested syntax and report back. karaf-maven-plugin does not merge bundle entry with URL parameters -- Key: KARAF-3286 URL: https://issues.apache.org/jira/browse/KARAF-3286 Project: Karaf Issue Type: Bug Components: karaf-tooling Affects Versions: 3.0.1 Reporter: Ralf Steppacher If a template contains an entry with URL parameters, e.g. {code:xml}bundlewrap:mvn:jboss/jbossall-client/4.2.3.GA/$Bundle-SymbolicName=jbossall-clientamp;Bundle-Version=4.2.3.GA/bundle{code} then it is not merged with a maven compile dependency. Instead the resulting feature.xml contains two entries for that bundle. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (KARAF-3286) karaf-maven-plugin does not merge bundle entry with URL parameters
[ https://issues.apache.org/jira/browse/KARAF-3286?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14172093#comment-14172093 ] Ralf Steppacher commented on KARAF-3286: No, it does not work with 3.0.2 either. The exact line that fails merging looks like this: {code:xml}bundlemvn:${project.groupId}/base/${project.version}/$Export-Package=an.api.package/bundle{code} The produced feature.xml contains the following two lines: {code:xml} bundlemvn:ch.vivates.ams/base/3.0.0-SNAPSHOT/$Export-Package=an.api.package/bundle bundlemvn:ch.vivates.ams/base/3.0.0-SNAPSHOT/bundle {code} karaf-maven-plugin does not merge bundle entry with URL parameters -- Key: KARAF-3286 URL: https://issues.apache.org/jira/browse/KARAF-3286 Project: Karaf Issue Type: Bug Components: karaf-tooling Affects Versions: 3.0.1 Reporter: Ralf Steppacher If a template contains an entry with URL parameters, e.g. {code:xml}bundlewrap:mvn:jboss/jbossall-client/4.2.3.GA/$Bundle-SymbolicName=jbossall-clientamp;Bundle-Version=4.2.3.GA/bundle{code} then it is not merged with a maven compile dependency. Instead the resulting feature.xml contains two entries for that bundle. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Created] (KARAF-3286) karaf-maven-plugin does not merge bundle entry with URL parameters
Ralf Steppacher created KARAF-3286: -- Summary: karaf-maven-plugin does not merge bundle entry with URL parameters Key: KARAF-3286 URL: https://issues.apache.org/jira/browse/KARAF-3286 Project: Karaf Issue Type: Bug Components: karaf-tooling Affects Versions: 3.0.1 Reporter: Ralf Steppacher If a template contains an entry with URL parameters, e.g. {code:xml}bundlewrap:mvn:jboss/jbossall-client/4.2.3.GA/$Bundle-SymbolicName=jbossall-clientamp;Bundle-Version=4.2.3.GA/bundle{code} then it is not merged with a maven compile dependency. Instead the resulting feature.xml contains two entries for that bundle. -- This message was sent by Atlassian JIRA (v6.3.4#6332)