Hi Ard, I apologize for the poor writing, it was very late at night here! Essentially the clustering function worked only on one event, the add of the document. Any subsequent changes to the first node were NOT relayed to the second node. Even this behavior does not happen consistently. Sometimes even the first event does not get relayed to the second node.
So it would seem that the configuration is valid if it worked at all. Just to provide a little more detail, I'm trying this with two Tomcat 6.x instances that are themselves clustered and then configuring JGroups to attach to the same IP but different ports. I've also tried configuring the two nodes to bind to separate interfaces and all attempts produced the same results. Thanks! -Eugene Ard Schrijvers wrote: > Aaah good point Nick....reading it again I read something else than I > did the first, though I must admit I thought I understood the issue the > first time, but now I have lost it: > > 'Subsequent modifications to the repo go unannounced to the second node' > > So, they go to the second node or not but unannounced? Or they are > unannounced, and don't go to the second node at all?? I am lost.. > > > >> I think the problem is in the last line: >> >> >>> Subsequent modifications to the repo go unannounced to the second >>> node. >>> >> With regards, >> >> Nick Stolwijk >> ~Java Developer~ >> >> Iprofs BV. >> Claus Sluterweg 125 >> 2012 WS Haarlem >> www.iprofs.nl >> >> >> >> On Fri, Nov 14, 2008 at 9:09 AM, Ard Schrijvers >> <[EMAIL PROTECTED]> wrote: >> >>> Hello Eugene >>> >>> >>>> Hi Folks, >>>> >>>> I'm getting a strange behavior after enabling the jgroups repo >>>> clustering. I see that both nodes "find" each other in the slide >>>> log. >>>> I then setup two tomcat instances (war builds), I add a >>>> >> document to >> >>>> one node and the second node gets the update. >>>> Subsequent modifications to the repo go unannounced to the second >>>> node. >>>> >>> Hmmm, I think I do not understand the problem...i suppose >>> >> if you add a >> >>> document to the first node, this node gets the new document >>> >> also, right? >> >>> So, but what is wrong in your opinion that the second node >>> >> also gets >> >>> the update? What do you expect from clustering? How it >>> >> normally works, >> >>> is that clustered repositories all write to the same >>> >> database, and the >> >>> notifications are send around just to invalidate the caches of the >>> other repo instances, such that, on a request, they >>> >> re-fetch the node >> >>> from the database. >>> >>> Aren't you confusing clustering with distributed storage? >>> >>> Ard >>> >>> >>>> I've definitely seen repo clustering working consistently at one >>>> point. >>>> Does this sound like anything obvious or familiar to anyone? >>>> >>>> TIA for any help! >>>> >>>> -- >>>> Eugene Fabrikant >>>> Bluenog Corporation >>>> >>>> bluenogef (Y!IM, AIM) >>>> >>>> >>>> >>>> ******************************************** >>>> Hippocms-dev: Hippo CMS development public mailinglist >>>> >>>> Searchable archives can be found at: >>>> MarkMail: http://hippocms-dev.markmail.org >>>> Nabble: http://www.nabble.com/Hippo-CMS-f26633.html >>>> >>>> >>>> >>> ******************************************** >>> Hippocms-dev: Hippo CMS development public mailinglist >>> >>> Searchable archives can be found at: >>> MarkMail: http://hippocms-dev.markmail.org >>> Nabble: http://www.nabble.com/Hippo-CMS-f26633.html >>> >>> >>> >> ******************************************** >> Hippocms-dev: Hippo CMS development public mailinglist >> >> Searchable archives can be found at: >> MarkMail: http://hippocms-dev.markmail.org >> Nabble: http://www.nabble.com/Hippo-CMS-f26633.html >> >> >> > ******************************************** > Hippocms-dev: Hippo CMS development public mailinglist > > Searchable archives can be found at: > MarkMail: http://hippocms-dev.markmail.org > Nabble: http://www.nabble.com/Hippo-CMS-f26633.html > > -- Eugene Fabrikant Bluenog Corporation bluenogef (Y!IM, AIM) ******************************************** Hippocms-dev: Hippo CMS development public mailinglist Searchable archives can be found at: MarkMail: http://hippocms-dev.markmail.org Nabble: http://www.nabble.com/Hippo-CMS-f26633.html
