Hi all, I think it is a good idea to release a new revision before adding a big changes to cluster management. This is a final release of 0.2.x. After adding a cluster driver framework, I'll release 0.3.0. This release includes bug fixes and some additional features: - support location-aware data placement We can specify the location of sheep daemons with a command line option '--zone'. Data is never replicated in the same zone. We can use this feature, for example, to replicate data to different racks by specifying the same zone in the same rack. - better cluster recovery Even if we stop sheep daemons without 'collie cluster shutdown', Sheepdog can reboot correctly now. (Of course, it is highly recommended to use 'collie cluster shutdown') - add new subcommands to collie Collie supports 'vdi create/snapshot/clone' (only qemu-img supported these operations before) and 'vdi read/write'. See also subcommand help: collie vdi [subcommand] -h - handle network partition failure Sheepdog kills minority nodes when a network partition has occurred. Though this approach kills many nodes, it is the most simple way to keep strong consistency. The software is available from GitHub: https://github.com/collie/sheepdog/archives/master Thanks, Kazutaka |