Lon Hohberger wrote: > On Wed, 2008-02-06 at 12:20 -0500, Lon Hohberger wrote: >> On Wed, 2008-02-06 at 10:21 -0500, Yan Vinogradov wrote: >> >>> Could be a bug. Does it say 'estranged' in the output, or not? >>> >>> -- Lon >>> >>> Hi Lon, yes, it does refer to the third node as estranged. Does it >>> tell you anything? >>> >> Clustat reports "Estranged" whenever it asks cman for a node list, and >> the corresponding node does not show up in cluster.conf. >> >> In your case, it means that cman still has state about the node despite >> the configuration change in ccsd. >> >> I'm not sure cman "purges" nodes that were dead (but part of >> cluster.conf) that later are removed from cluster.conf. >> >> Chrissie might know. That's right, cman never removes a node from its list. That's chiefly because it can't easily tell whether it might suddenly appear again, which was mainly the case in RHEL4 where cman was in the kernel and didn't have easy access to cluster.conf. Actually, in RHEL5 it could be possibly to purge a node from cman, it is something I've considered and might have a look at in the future, if people think it's worthwhile. Chrissie -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
