Evening, I have been looking at MAAS as a possible solution to condense several of our existing clusters into being more efficient. We already running anything from several dozen-node Cassandra clusters to several hundreds of nodes running Hadoop.
My Google-fu seems very inadequate at this point because I have been looking at what the recommended hardware configuration for nodes should be. Some questions: * Should the node use RAID? or JBOD for the drives? Will it be configured properly if you use JBOD? - We use JBOD for all of our clusters to spread out the disk IO on as many spindles as possible. We rely on the replication as a means of data reliance. * If you have multiple ethernet interfaces is there a preferred configuration for that? Can the system use multiple ethernet ports for the purpose of bonding or increasing bandwidth? - We use at least two GigE ports for our slow clusters and 10g for all of the heavy use clusters. If we have the extra interfaces we usually also have an administrative network interface independent of the data network. I guess that is a good start for now. Thanks. Richard -- ubuntu-server mailing list ubuntu-server@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-server More info: https://wiki.ubuntu.com/ServerTeam