ceph block size
hi, when an application within a VM write to ceph cluster then every block written is broken by ceph (OSD) in smaller parts and is written -stripe accross the ceph nodes. right? by doing this the read and write will be faster than using normal block. right? tnx
It's not really faster because you have the replication, network latency, and some more limitations. But as @Bernd already states, the design is built for reliability so your data is still available even if some of your hard disks fail.
I forgot to mention that you don't have a SOP since every client talks to his "own" OSD daemons directly, so it may not be faster, but the cluster handles way more requests in parallel.
Yes, whereas a single application won’t get performance benefits out of Ceph, the total performance of a Ceph cluster goes up as it scales up. In other words, as it scales it can handle more load.
every i/o is broken down in blocks and each block is written in a ceph node.this block will be then copied to the rest of the ceph node?
I don't know whether it is done at the block level, but a client writes data to an OSD, which then replicates it to the other OSDs that contain copies of that data. See http://docs.ceph.com/docs/giant/archi..., point 4 "replication".