[Opensymphony-oscache] More clustering on 2.1RC1
Jonas Larsson <[email protected]> Tue, 24 May 2005 09:23:35 CDT
| Newsgroups | gmane.comp.java.open-symphony.os-cache |
|---|---|
| Message-ID | <33207312.1116944712231.JavaMail.os-j2ee@opensymphony01.contegix.com> |
I get a bit worried when I see this on application startup in my logs when testing on my development cluster (two-instance Weblogic 8.1SP3 cluster, one admin server and two managed servers): [05/24/05 13:40:44.683 INFO ] A new member at address 'lonlxwebhost3:58488' has joined the cluster (JavaGroupsBroadcastingListener.java:166) in thread "PullPushAdapterThread" [05/24/05 13:40:44.686 INFO ] A new member at address 'lonlxwebhost6:33524' has joined the cluster (JavaGroupsBroadcastingListener.java:166) in thread "PullPushAdapterThread" [05/24/05 13:40:44.689 INFO ] A new member at address 'lonlxwebhost9:53696' has joined the cluster (JavaGroupsBroadcastingListener.java:166) in thread "PullPushAdapterThread" [05/24/05 13:40:44.690 INFO ] A new member at address 'lonlxwebhost10:57908' has joined the cluster (JavaGroupsBroadcastingListener.java:166) in thread "PullPushAdapterThread" [05/24/05 13:40:44.690 INFO ] A new member at address 'lonlxwebhost10:62332' has joined the cluster (JavaGroupsBroadcastingListener.java:166) in thread "PullPushAdapterThread" [05/24/05 13:43:34.020 INFO ] A new member at address 'lonlxwebhost3:33777' has joined the cluster (JavaGroupsBroadcastingListener.java:166) in thread "PullPushAdapterThread" This is from the log of one of the instances in my development cluster (on 3 and 10). What makes me a bit nervous is that the production (6 and 9) and staging (also on 3 and 10 but different ports) instances also seem to want to play even though the OSCache clustering docs says that if one wants to run multiple deployments of the same application on a LAN just use different multicast addresses (one for Prod, one for Stage and one for Dev). I have gone through great pain to aquire these additional multicast addresses (working on a big company is not always easy) and it still seems to want to hook up to all instances I have running... It still seems to work fine since we only flush once at end of day but it would be nice if someone could help me verify how this is supposed to work. The SA's is going to kill me if we have rougue multicast applications that is sending stuff all over the place... --------------------------------------------------------------------- Posted via Jive Forums http://forums.opensymphony.com/thread.jspa?threadID=2503&messageID=7135#7135