Try this: - service gfs stop - if you use shared storage(or gfs2) - service rgmanager stop - service clvmd stop - service cman stop It usually works for me on a 2 node Centos5 cluster. ---- JCascao > -----Original Message----- > From: linux-cluster-bounces@xxxxxxxxxx > [mailto:linux-cluster-bounces@xxxxxxxxxx]On Behalf Of James Fidell > Sent: terça-feira, 6 de Novembro de 2007 15:02 > To: linux clustering > Subject: Re: Shutting down a cluster > > > Patrick Caulfield wrote: > > Ward, Timothy - SSD wrote: > >> Try: > >> # cman_tool force leave > >> > > > > I would try and avoid that if possible. It's much better to > make sure that all > > of the services are shut down beforehand and let cman shut > down tidily. > > > > Check for things like ccsd still running and services shown > my 'cman_tool > > services'. Unfortunately it can be hard to determine which > cluster services are > > running if you don't know what they all are, but it's worth > trying :-) > > Ok. I have a newly rebooted server on which I've started clustering, > but not started any services (other than those started by the > cman init > script). > > # cman_tool services > > reports that fenced is the only service running. So, I killed fenced. > Nothing else appears to notice that fenced on this server has > died. Now > > # cman_tool leave remove > > works, in that it returns no errors. But within a few seconds, the > leaving node is fenced and shut down :( > > James > > -- > Linux-cluster mailing list > Linux-cluster@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/linux-cluster > > !DSPAM:3,473081ef76021170020771! > -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
