Is a master-to-multiple-slave setup functionally the same as "device ganging" per the H2 design document? Would such a setup mean that a failed drive could be rebuilt by adding a new block device, designating it as either a master or slave and the data would be rebuilt on that drive? Or said in another way, can we finally get rid of the RAID cards if our main aim is automatic fail-over?
M On 07/22/2018 09:20 PM, Matthew Dillon wrote:
Multi-master and multi-volume will take more time, but I think I will be able to get master-to-multiple-slave operation on H2 working soon and that will be roughly as efficient as mirror-stream is for H1. There's still a ton of work that needs to be done on H2, but single-storage support is rock solid and H2's compression features are really nice. H2 also has working snapshot support, but it isn't automatic like it is on H1. H2 snapshots are writable though, while H1 snapshots are read-only. -Matt
