[xwiki-users] What can go wrong when setting up clustering?
Hello XWikiers, I've been following the tutorial: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/DistributedEventClusterS... and the two consoles show me a successful: grep -i channel mytomcat/logs/tomcat.log
2012-07-06 08:28:02,901 INFO jgroups.JChannel - JGroups version: 2.12.2.Final 2012-07-06 08:28:05,262 [main] INFO .o.r.i.j.JGroupsNetworkAdapter - Channel [udp] started
grep -i channel myothertomcat/logs/tomcat.log
2012-07-06 08:29:11,857 INFO jgroups.JChannel - JGroups version: 2.12.2.Final 2012-07-06 08:29:14,251 [main] INFO .o.r.i.j.JGroupsNetworkAdapter - Channel [udp] started
also; GMS: address=atlas-56733, cluster=event, physical address=fe80:0:0:0:223:dfff:fedf:591d%5:2273 GMS: address=atlas-26026, cluster=event, physical address=fe80:0:0:0:223:dfff:fedf:591d%5:20950 with unchanged udp.xml as extracted with jar -xf from the jgroups jar in WEB-INF. However the two xwiki's seem to work separately, the data of the other is only fetched if it has not yet come into cache. The only suspicious log is, in both consoles:
2012-07-06 08:28:03,208 WARN protocols.UDP - receive buffer of socket java.net.MulticastSocket@40d3ab8b was set to 25MB, but the OS only allocated 65.51KB. This might lead to performance problems. Please set your max receive buffer in the OS correctly (e.g. net.core.rmem_max on Linux)
Is there anything else I need to watch? The README in the directory of udp.xml was absolutely useless, or at least lead me to no change... have badly read it? thanks in advance Paul
On Fri, Jul 6, 2012 at 1:22 PM, Paul Libbrecht <[email protected]> wrote:
Hello XWikiers,
I've been following the tutorial: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/DistributedEventClusterS... and the two consoles show me a successful:
grep -i channel mytomcat/logs/tomcat.log
2012-07-06 08:28:02,901 INFO jgroups.JChannel - JGroups version: 2.12.2.Final 2012-07-06 08:28:05,262 [main] INFO .o.r.i.j.JGroupsNetworkAdapter - Channel [udp] started
grep -i channel myothertomcat/logs/tomcat.log
2012-07-06 08:29:11,857 INFO jgroups.JChannel - JGroups version: 2.12.2.Final 2012-07-06 08:29:14,251 [main] INFO .o.r.i.j.JGroupsNetworkAdapter - Channel [udp] started
also; GMS: address=atlas-56733, cluster=event, physical address=fe80:0:0:0:223:dfff:fedf:591d%5:2273
GMS: address=atlas-26026, cluster=event, physical address=fe80:0:0:0:223:dfff:fedf:591d%5:20950
with unchanged udp.xml as extracted with jar -xf from the jgroups jar in WEB-INF.
However the two xwiki's seem to work separately, the data of the other is only fetched if it has not yet come into cache.
The only suspicious log is, in both consoles:
2012-07-06 08:28:03,208 WARN protocols.UDP - receive buffer of socket java.net.MulticastSocket@40d3ab8b was set to 25MB, but the OS only allocated 65.51KB. This might lead to performance problems. Please set your max receive buffer in the OS correctly (e.g. net.core.rmem_max on Linux)
Is there anything else I need to watch? The README in the directory of udp.xml was absolutely useless, or at least lead me to no change... have badly read it?
You could enable debug log to see if cluster members talk to each others. See http://extensions.xwiki.org/xwiki/bin/view/Extension/Observation+Module+Remo....
thanks in advance
Paul _______________________________________________ users mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/users
-- Thomas Mortagne
Hi, On Jul 6, 2012, at 1:22 PM, Paul Libbrecht wrote:
Hello XWikiers,
I've been following the tutorial: http://platform.xwiki.org/xwiki/bin/view/AdminGuide/DistributedEventClusterS... and the two consoles show me a successful:
grep -i channel mytomcat/logs/tomcat.log
2012-07-06 08:28:02,901 INFO jgroups.JChannel - JGroups version: 2.12.2.Final 2012-07-06 08:28:05,262 [main] INFO .o.r.i.j.JGroupsNetworkAdapter - Channel [udp] started
grep -i channel myothertomcat/logs/tomcat.log
2012-07-06 08:29:11,857 INFO jgroups.JChannel - JGroups version: 2.12.2.Final 2012-07-06 08:29:14,251 [main] INFO .o.r.i.j.JGroupsNetworkAdapter - Channel [udp] started
also; GMS: address=atlas-56733, cluster=event, physical address=fe80:0:0:0:223:dfff:fedf:591d%5:2273
GMS: address=atlas-26026, cluster=event, physical address=fe80:0:0:0:223:dfff:fedf:591d%5:20950
with unchanged udp.xml as extracted with jar -xf from the jgroups jar in WEB-INF.
However the two xwiki's seem to work separately, the data of the other is only fetched if it has not yet come into cache.
The only suspicious log is, in both consoles:
2012-07-06 08:28:03,208 WARN protocols.UDP - receive buffer of socket java.net.MulticastSocket@40d3ab8b was set to 25MB, but the OS only allocated 65.51KB. This might lead to performance problems. Please set your max receive buffer in the OS correctly (e.g. net.core.rmem_max on Linux)
Is there anything else I need to watch? The README in the directory of udp.xml was absolutely useless, or at least lead me to no change... have badly read it?
You could check with JMX to see the status of the cluster, see http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Monitoring#HJGroupsMonit... Thanks -Vincent
Is there anything else I need to watch?
Le 6 juil. 2012 à 13:28, Thomas Mortagne a écrit :
You could enable debug log to see if cluster members talk to each others. See http://extensions.xwiki.org/xwiki/bin/view/Extension/Observation+Module+Remo....
I did this: into WEB-INF/classes/logback.xml, added: <!-- remote observation detailed logging --> <logger name="org.xwiki.observation.remote" level="trace"/> <logger name="org.jgroups" level="trace"/> and on the side of the sender I did get the nice following debug line: 2012-07-06 17:02:28,122 [http://hoplahup.homeip.net/xwiki/bin/saveandcontinue/Main/Test3] DEBUG .o.r.i.j.JGroupsNetworkAdapter - Send JGroups remote event [event: [org.xwiki.bridge.event.DocumentUpdatedEvent@b6362ec], source: [{origdoclanguage=, origdocversion=8.1, docversion=9.1, doclanguage=, docname=name = [Test3], type = [DOCUMENT], parent = [name = [Main], type = [SPACE], parent = [name = [xwiki], type = [WIKI], parent = [null]]]}], data: [{contextuser=XWiki.adminPolx, contextwiki=xwiki}]] But not much more. However, I do not seem to see a connection to the other "node". Le 6 juil. 2012 à 13:30, Vincent Massol a écrit :
You could check with JMX to see the status of the cluster, see http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Monitoring#HJGroupsMonit...
Vincent, that seems like something very useful but how do you read the connected clients there? Which attribute would list this? I see neither on your screenshot or my jconsole. thanks again! Paul
After having moved to the tcp-nio protocol stack, and manually adjusting the port numbers for each instance, I could finally get it to work. I am still having announces that some ports are connection refused but these are not the ports I configured but the port +1... is that classical? Is there an advice for us to use UDP? I would have expected this to be worry free on a simple desktop when looking at localhost... but well. Paul Le 6 juil. 2012 à 18:14, Paul Libbrecht a écrit :
Is there anything else I need to watch?
Le 6 juil. 2012 à 13:28, Thomas Mortagne a écrit :
You could enable debug log to see if cluster members talk to each others. See http://extensions.xwiki.org/xwiki/bin/view/Extension/Observation+Module+Remo....
I did this: into WEB-INF/classes/logback.xml, added: <!-- remote observation detailed logging --> <logger name="org.xwiki.observation.remote" level="trace"/> <logger name="org.jgroups" level="trace"/>
and on the side of the sender I did get the nice following debug line:
2012-07-06 17:02:28,122 [http://hoplahup.homeip.net/xwiki/bin/saveandcontinue/Main/Test3] DEBUG .o.r.i.j.JGroupsNetworkAdapter - Send JGroups remote event [event: [org.xwiki.bridge.event.DocumentUpdatedEvent@b6362ec], source: [{origdoclanguage=, origdocversion=8.1, docversion=9.1, doclanguage=, docname=name = [Test3], type = [DOCUMENT], parent = [name = [Main], type = [SPACE], parent = [name = [xwiki], type = [WIKI], parent = [null]]]}], data: [{contextuser=XWiki.adminPolx, contextwiki=xwiki}]]
But not much more.
However, I do not seem to see a connection to the other "node".
Le 6 juil. 2012 à 13:30, Vincent Massol a écrit :
You could check with JMX to see the status of the cluster, see http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Monitoring#HJGroupsMonit...
Vincent, that seems like something very useful but how do you read the connected clients there? Which attribute would list this? I see neither on your screenshot or my jconsole.
On Sat, Jul 7, 2012 at 1:04 AM, Paul Libbrecht <[email protected]> wrote:
After having moved to the tcp-nio protocol stack, and manually adjusting the port numbers for each instance, I could finally get it to work.
I am still having announces that some ports are connection refused but these are not the ports I configured but the port +1... is that classical?
Is there an advice for us to use UDP? I would have expected this to be worry free on a simple desktop when looking at localhost... but well.
Sorry Paul but we are not yet JGroups experts ourself, I mostly used TCP for tests. You might get more valuable informations from JGroups mailing list.
Paul
Le 6 juil. 2012 à 18:14, Paul Libbrecht a écrit :
Is there anything else I need to watch?
Le 6 juil. 2012 à 13:28, Thomas Mortagne a écrit :
You could enable debug log to see if cluster members talk to each others. See http://extensions.xwiki.org/xwiki/bin/view/Extension/Observation+Module+Remo....
I did this: into WEB-INF/classes/logback.xml, added: <!-- remote observation detailed logging --> <logger name="org.xwiki.observation.remote" level="trace"/> <logger name="org.jgroups" level="trace"/>
and on the side of the sender I did get the nice following debug line:
2012-07-06 17:02:28,122 [http://hoplahup.homeip.net/xwiki/bin/saveandcontinue/Main/Test3] DEBUG .o.r.i.j.JGroupsNetworkAdapter - Send JGroups remote event [event: [org.xwiki.bridge.event.DocumentUpdatedEvent@b6362ec], source: [{origdoclanguage=, origdocversion=8.1, docversion=9.1, doclanguage=, docname=name = [Test3], type = [DOCUMENT], parent = [name = [Main], type = [SPACE], parent = [name = [xwiki], type = [WIKI], parent = [null]]]}], data: [{contextuser=XWiki.adminPolx, contextwiki=xwiki}]]
But not much more.
However, I do not seem to see a connection to the other "node".
Le 6 juil. 2012 à 13:30, Vincent Massol a écrit :
You could check with JMX to see the status of the cluster, see http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Monitoring#HJGroupsMonit...
Vincent, that seems like something very useful but how do you read the connected clients there? Which attribute would list this? I see neither on your screenshot or my jconsole.
_______________________________________________ users mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/users
-- Thomas Mortagne
Hi, The tcp-nio protocol is labeled as experimental so there might be surprises with that as opposed to tcp. The udp protocol depends on multicast which is likely to be broken in real world networks as almost nothing depends on it. I observed the pinging of the next higher port in my testing, here in the default tcp.xml it is configured to ping 7800 and 7801 https://github.com/belaban/JGroups/blob/master/conf/tcp.xml#L45 Caleb On 07/07/2012 03:49 AM, Thomas Mortagne wrote:
On Sat, Jul 7, 2012 at 1:04 AM, Paul Libbrecht <[email protected]> wrote:
After having moved to the tcp-nio protocol stack, and manually adjusting the port numbers for each instance, I could finally get it to work.
I am still having announces that some ports are connection refused but these are not the ports I configured but the port +1... is that classical?
Is there an advice for us to use UDP? I would have expected this to be worry free on a simple desktop when looking at localhost... but well.
Sorry Paul but we are not yet JGroups experts ourself, I mostly used TCP for tests. You might get more valuable informations from JGroups mailing list.
Paul
Le 6 juil. 2012 à 18:14, Paul Libbrecht a écrit :
Is there anything else I need to watch?
Le 6 juil. 2012 à 13:28, Thomas Mortagne a écrit :
You could enable debug log to see if cluster members talk to each others. See http://extensions.xwiki.org/xwiki/bin/view/Extension/Observation+Module+Remo....
I did this: into WEB-INF/classes/logback.xml, added: <!-- remote observation detailed logging --> <logger name="org.xwiki.observation.remote" level="trace"/> <logger name="org.jgroups" level="trace"/>
and on the side of the sender I did get the nice following debug line:
2012-07-06 17:02:28,122 [http://hoplahup.homeip.net/xwiki/bin/saveandcontinue/Main/Test3] DEBUG .o.r.i.j.JGroupsNetworkAdapter - Send JGroups remote event [event: [org.xwiki.bridge.event.DocumentUpdatedEvent@b6362ec], source: [{origdoclanguage=, origdocversion=8.1, docversion=9.1, doclanguage=, docname=name = [Test3], type = [DOCUMENT], parent = [name = [Main], type = [SPACE], parent = [name = [xwiki], type = [WIKI], parent = [null]]]}], data: [{contextuser=XWiki.adminPolx, contextwiki=xwiki}]]
But not much more.
However, I do not seem to see a connection to the other "node".
Le 6 juil. 2012 à 13:30, Vincent Massol a écrit :
You could check with JMX to see the status of the cluster, see http://platform.xwiki.org/xwiki/bin/view/AdminGuide/Monitoring#HJGroupsMonit...
Vincent, that seems like something very useful but how do you read the connected clients there? Which attribute would list this? I see neither on your screenshot or my jconsole.
_______________________________________________ users mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/users
Le 8 juil. 2012 à 08:44, Caleb James DeLisle a écrit :
The tcp-nio protocol is labeled as experimental so there might be surprises with that as opposed to tcp.
thanks, that had escaped me. I have now moved to TCP.
The udp protocol depends on multicast which is likely to be broken in real world networks as almost nothing depends on it. I observed the pinging of the next higher port in my testing, here in the default tcp.xml it is configured to ping 7800 and 7801 https://github.com/belaban/JGroups/blob/master/conf/tcp.xml#L45
Ok, now I have something like a working jgroups config. My two hosts are called dev-current-app and dev-next-app, they are configured together with both tcp.xml configured to look for each other. I can see the following in my logs after a document save (I save an xwiki object): [#|2012-07-18T08:56:02.726-0700|INFO|sun-appserver9.1|javax.enterprise.system.stream.out|_ThreadID=24;_ThreadName=http://current.dev.curriki.org/xwiki/bin/view/CurrikiCode/AssetSaveService;|2012-07-18 08:56:02,725 [http://current.dev.curriki.org/xwiki/bin/view/CurrikiCode/AssetSaveService] DEBUG .o.r.i.j.JGroupsNetworkAdapter - Send JGroups remote event [event: [org.xwiki.bridge.event.DocumentUpdatedEvent@5b91ebe], source: [{origdoclanguage=, origdocversion=11.1, docversion=12.1, doclanguage=, docname=name = [testohtest_0], type = [DOCUMENT], parent = [name = [Coll_polx], type = [SPACE], parent = [name = [xwiki], type = [WIKI], parent = [null]]]}], data: [{contextuser=XWiki.polx, contextwiki=xwiki}]] [#|2012-07-18T08:56:02.604-0700|INFO|sun-appserver9.1|javax.enterprise.system.stream.out|_ThreadID=37;_ThreadName=Incoming-3,event,dev-next-app-26832;|2012-07-18 08:56:02,604 [Incoming-3,event,dev-next-app-26832] DEBUG o.r.i.j.DefaultJGroupsReceiver - Received JGroups remote event [event: [org.xwiki.bridge.event.DocumentUpdatedEvent@5b91ebe], source: [{origdocversion=11.1, origdoclanguage=, docversion=12.1, doclanguage=, docname=name = [testohtest_0], type = [DOCUMENT], parent = [name = [Coll_polx], type = [SPACE], parent = [name = [xwiki], type = [WIKI], parent = [null]]]}], data: [{contextwiki=xwiki, contextuser=XWiki.polx}]] and I see no exception reported (and a zillion "pings"). nonetheless, the change i made on current-dev only comes to next-dev if I flush cache which, we agree, is not the solution. Where else could I look for a bug? thanks in advance. Paul
Hello XWiki experts, ...and a bit later, everything sounds to work very well between next.dev.curriki.org and current.dev.curriki.org except one thing: obtaining older versions from the Hibernate store within the notifications interface fails. This happens when a change of document is made in one node and the listener is running on another node.
[#|2012-07-18T07:27:45.976-0700|WARNING|sun-appserver9.1|javax.enterprise.system.stream.err|_ThreadID=58;_ThreadName=Incoming-9,event,Dev-current-app-33350;_Reques tID=89a8c4db-c4b5-4192-a6e4-fade41ae14ba;| Comparing version : 10.1 to 9.1|#]
[#|2012-07-18T07:27:46.008-0700|INFO|sun-appserver9.1|javax.enterprise.system.stream.out|_ThreadID=58;_ThreadName=Incoming-9,event,Dev-current-app-33350;|2012-07-1 8 07:27:46,006 [Incoming-9,event,Dev-current-app-33350] WARN g.S.DocChangeListener - Trouble for indexing Coll_polx.testohtest_0 com.xpn.xwiki.XWikiException: Error number 3205 in 3: Version 9.1 does not exist while reading document Coll_polx.testohtest_0 at com.xpn.xwiki.store.XWikiHibernateVersioningStore.loadXWikiDoc(XWikiHibernateVersioningStore.java:216) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.XWiki.getDocument(XWiki.java:1448) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.doc.XWikiDocument.getMetaDataDiff(XWikiDocument.java:5454) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.doc.XWikiDocument$getMetaDataDiff$0.call(Unknown Source) ~[na:na] at org.codehaus.groovy.runtime.callsite.CallSiteArray.defaultCall(CallSiteArray.java:42) ~[groovy-all-1.8.6.jar:1.8.6] at org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:108) ~[groovy-all-1.8.6.jar:1.8.6] at org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:120) ~[groovy-all-1.8.6.jar:1.8.6] at DocChangeListener.notify(script1342563833153805798160.groovy:100) ~[na:na] at com.xpn.xwiki.notify.DocChangeRule.verify(DocChangeRule.java:78) [xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.notify.XWikiNotificationManager.verify(XWikiNotificationManager.java:208) [xwiki-platform-legacy-oldcore-3.5.jar:na] at org.xwiki.legacy.internal.oldcore.notification.LegacyNotificationDispatcher.onEvent(LegacyNotificationDispatcher.java:115) [xwiki-platform-legacy-oldcor e-3.5.jar:na] at org.xwiki.observation.internal.DefaultObservationManager.notify(DefaultObservationManager.java:268) [xwiki-commons-observation-local-3.5.jar:na] at org.xwiki.observation.internal.DefaultObservationManager.notify(DefaultObservationManager.java:233) [xwiki-commons-observation-local-3.5.jar:na]
Is clustering incompatible to "legacy-oldcore" ? thanks in advance Paul
Well, there can be another interpretation to the exception I report below: the jgroup even notifies the new version has occurred (and has arrived) but probably the previous would not have arrived yet. Does it make sense? Do I have another way to tell that the notification has arrived from a remote node of the cluster? (we are still using the old notification mechanism but maybe this is a good reason to upgrade). thank you much Paul Le 20 juil. 2012 à 21:58, Paul Libbrecht a écrit :
Hello XWiki experts,
...and a bit later, everything sounds to work very well between next.dev.curriki.org and current.dev.curriki.org except one thing: obtaining older versions from the Hibernate store within the notifications interface fails. This happens when a change of document is made in one node and the listener is running on another node.
[#|2012-07-18T07:27:45.976-0700|WARNING|sun-appserver9.1|javax.enterprise.system.stream.err|_ThreadID=58;_ThreadName=Incoming-9,event,Dev-current-app-33350;_Reques tID=89a8c4db-c4b5-4192-a6e4-fade41ae14ba;| Comparing version : 10.1 to 9.1|#]
[#|2012-07-18T07:27:46.008-0700|INFO|sun-appserver9.1|javax.enterprise.system.stream.out|_ThreadID=58;_ThreadName=Incoming-9,event,Dev-current-app-33350;|2012-07-1 8 07:27:46,006 [Incoming-9,event,Dev-current-app-33350] WARN g.S.DocChangeListener - Trouble for indexing Coll_polx.testohtest_0 com.xpn.xwiki.XWikiException: Error number 3205 in 3: Version 9.1 does not exist while reading document Coll_polx.testohtest_0 at com.xpn.xwiki.store.XWikiHibernateVersioningStore.loadXWikiDoc(XWikiHibernateVersioningStore.java:216) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.XWiki.getDocument(XWiki.java:1448) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.doc.XWikiDocument.getMetaDataDiff(XWikiDocument.java:5454) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.doc.XWikiDocument$getMetaDataDiff$0.call(Unknown Source) ~[na:na] at org.codehaus.groovy.runtime.callsite.CallSiteArray.defaultCall(CallSiteArray.java:42) ~[groovy-all-1.8.6.jar:1.8.6] at org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:108) ~[groovy-all-1.8.6.jar:1.8.6] at org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:120) ~[groovy-all-1.8.6.jar:1.8.6] at DocChangeListener.notify(script1342563833153805798160.groovy:100) ~[na:na] at com.xpn.xwiki.notify.DocChangeRule.verify(DocChangeRule.java:78) [xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.notify.XWikiNotificationManager.verify(XWikiNotificationManager.java:208) [xwiki-platform-legacy-oldcore-3.5.jar:na] at org.xwiki.legacy.internal.oldcore.notification.LegacyNotificationDispatcher.onEvent(LegacyNotificationDispatcher.java:115) [xwiki-platform-legacy-oldcor e-3.5.jar:na] at org.xwiki.observation.internal.DefaultObservationManager.notify(DefaultObservationManager.java:268) [xwiki-commons-observation-local-3.5.jar:na] at org.xwiki.observation.internal.DefaultObservationManager.notify(DefaultObservationManager.java:233) [xwiki-commons-observation-local-3.5.jar:na]
Is clustering incompatible to "legacy-oldcore" ?
thanks in advance
Paul _______________________________________________ users mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/users
On Wed, Jul 25, 2012 at 11:14 PM, Paul Libbrecht <[email protected]> wrote:
Well, there can be another interpretation to the exception I report below: the jgroup even notifies the new version has occurred (and has arrived) but probably the previous would not have arrived yet.
Does it make sense?
Do I have another way to tell that the notification has arrived from a remote node of the cluster?
You have the information in the execution context. You can look at how cache store makes the difference for example. See https://github.com/xwiki/xwiki-platform/blob/master/xwiki-platform-core/xwik...
(we are still using the old notification mechanism but maybe this is a good reason to upgrade).
thank you much
Paul
Le 20 juil. 2012 à 21:58, Paul Libbrecht a écrit :
Hello XWiki experts,
...and a bit later, everything sounds to work very well between next.dev.curriki.org and current.dev.curriki.org except one thing: obtaining older versions from the Hibernate store within the notifications interface fails. This happens when a change of document is made in one node and the listener is running on another node.
[#|2012-07-18T07:27:45.976-0700|WARNING|sun-appserver9.1|javax.enterprise.system.stream.err|_ThreadID=58;_ThreadName=Incoming-9,event,Dev-current-app-33350;_Reques tID=89a8c4db-c4b5-4192-a6e4-fade41ae14ba;| Comparing version : 10.1 to 9.1|#]
[#|2012-07-18T07:27:46.008-0700|INFO|sun-appserver9.1|javax.enterprise.system.stream.out|_ThreadID=58;_ThreadName=Incoming-9,event,Dev-current-app-33350;|2012-07-1 8 07:27:46,006 [Incoming-9,event,Dev-current-app-33350] WARN g.S.DocChangeListener - Trouble for indexing Coll_polx.testohtest_0 com.xpn.xwiki.XWikiException: Error number 3205 in 3: Version 9.1 does not exist while reading document Coll_polx.testohtest_0 at com.xpn.xwiki.store.XWikiHibernateVersioningStore.loadXWikiDoc(XWikiHibernateVersioningStore.java:216) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.XWiki.getDocument(XWiki.java:1448) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.doc.XWikiDocument.getMetaDataDiff(XWikiDocument.java:5454) ~[xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.doc.XWikiDocument$getMetaDataDiff$0.call(Unknown Source) ~[na:na] at org.codehaus.groovy.runtime.callsite.CallSiteArray.defaultCall(CallSiteArray.java:42) ~[groovy-all-1.8.6.jar:1.8.6] at org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:108) ~[groovy-all-1.8.6.jar:1.8.6] at org.codehaus.groovy.runtime.callsite.AbstractCallSite.call(AbstractCallSite.java:120) ~[groovy-all-1.8.6.jar:1.8.6] at DocChangeListener.notify(script1342563833153805798160.groovy:100) ~[na:na] at com.xpn.xwiki.notify.DocChangeRule.verify(DocChangeRule.java:78) [xwiki-platform-legacy-oldcore-3.5.jar:na] at com.xpn.xwiki.notify.XWikiNotificationManager.verify(XWikiNotificationManager.java:208) [xwiki-platform-legacy-oldcore-3.5.jar:na] at org.xwiki.legacy.internal.oldcore.notification.LegacyNotificationDispatcher.onEvent(LegacyNotificationDispatcher.java:115) [xwiki-platform-legacy-oldcor e-3.5.jar:na] at org.xwiki.observation.internal.DefaultObservationManager.notify(DefaultObservationManager.java:268) [xwiki-commons-observation-local-3.5.jar:na] at org.xwiki.observation.internal.DefaultObservationManager.notify(DefaultObservationManager.java:233) [xwiki-commons-observation-local-3.5.jar:na]
Is clustering incompatible to "legacy-oldcore" ?
thanks in advance
Paul _______________________________________________ users mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/users
_______________________________________________ users mailing list [email protected] http://lists.xwiki.org/mailman/listinfo/users
-- Thomas Mortagne
participants (4)
-
Caleb James DeLisle -
Paul Libbrecht -
Thomas Mortagne -
Vincent Massol