Michael Will wrote: >> If you stick to the numerical cluster I'd and just use a more robust >> hash alg you should be fine... >> > We can't change the hash algorithm as it would break backwards compatibility - > new nodes would hash the same name to a different > value and refuse to join the cluster. > A better solution might be to allow the cluster number to be specified in ccs > as well as the name so it would override the hash value. > -- patrick Yes, this seems also to me a good solution: if a new field "cluster_id" is set in cluster.conf (together with "cluster_name") generate_cluster_id takes it, and if not, it generates the id as usual. Is there any possibility for such an evolution in a future patch on CS4 U4 ? Thanks Alain -- mailto:Alain.Moulle@xxxxxxxx +------------------------------+--------------------------------+ | Alain Moullé | from France : 04 76 29 75 99 | | | FAX number : 04 76 29 72 49 | | Bull SA | | | 1, Rue de Provence | Adr : FREC B1-041 | | B.P. 208 | | | 38432 Echirolles - CEDEX | Email: Alain.Moulle@xxxxxxxx | | France | BCOM : 229 7599 | +-------------------------------+-------------------------------+ -- Linux-cluster mailing list Linux-cluster@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/linux-cluster
