how does flockdb workwith gizzard #13

Open
crazyfox3 opened this Issue Sep 11, 2010 · 3 comments

Projects

None yet

2 participants

@crazyfox3

I really can not get it about how does flockdb work with multi pcserver, I mean how does flockdb spread the data to multi pcserver, is every shard have the same data or different with others? how does gizzard work in flockdb?

thanks a lot , i really confused about it thoes day....

@robey
robey commented Sep 13, 2010

flockdb uses gizzard for sharding. gizzard maps ints to shards using the forwarding table. since flockdb is using ints for source/destination id, it can just use those ints directly for the shard mapping.

@crazyfox3

yeah, ok, I probable understand what you mean. thank you. but where do those ints save, in mysql database or a file or the source code. And i find only one place in the config file (production.conf) where could configure the ip address, but i think that is for relicas, is that right?

@robey
robey commented Sep 14, 2010

the name server replicas are used to store the forwarding tables that map ids to shard trees.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment