Farming hangs

trossmy Sep 12, 2005 11:23 AM

Hi!
This only happens if on Linux:
Cluster with Debian 1:3.3.5-12
jboss 4.0.2
If I close down the coordinator all the other nodes leave and join a view forever like this:

2005-08-26 18:04:25,956 INFO [org.jboss.cache.TreeCache] viewAccepted(): new members: [mcm3:45360, xxx.xxx.xxx.xxx:33130]
2005-08-26 18:04:25,962 WARN [org.jgroups.protocols.pbcast.GMS] checkSelfInclusion() failed, xxx.xxx.xxx.xxx:33130 is not a member of view [mcm3:45360|136] [mcm3:45360]; discarding view
2005-08-26 18:04:25,963 WARN [org.jgroups.protocols.pbcast.GMS] I (xxx.xxx.xxx.xxx:33130) am being shunned, will leave and rejoin group (prev_members are [mcm3:45344 mcm3:45360 xxx.xxx.xxx.xxx:33130 ])

Anybody got a clue?
Also recently if I redeploy an archive in the farm directory, the application in question gets undeployed on all the nodes, but will only be redeployed on the coordinator?

1. Re: Farming hangs

trossmy Sep 14, 2005 7:41 AM (in response to trossmy)

I am happy to say that updating to jgroups 2.2.8 seems to have solved the issue. Note that you have to completely restart the whole cluster as different versions of jgroups don't like to talk with one another
Actions