Chris Lu
51939efeac
1. volume server now sends master server its max file key, so that
master server does not need to store the sequence on disk any more
2. fix raft server's failure to init cluster during bootstrapping
11 years ago
Chris Lu
cbc5a76e80
Added batch file deleting.
11 years ago
Chris Lu
460923f094
use fileId as the official name for the file identifier
11 years ago
Chris Lu
f20ef922fd
1. add batched volume lookup handler
2. working-in-progress batch delete
11 years ago
Chris Lu
f7f582ec86
1. refactoring, merge "replication" logic into "topology" package
2. when growing volumes, additional preferred "rack" and "dataNode"
paraemters are also provided. Previously only "dataCenter" paraemter is
provided.
11 years ago
Chris Lu
008aee0dc1
Add retrying logic to wait for other peers during cluster bootstrapping.
11 years ago
Chris Lu
39b774a131
1. adding statistics reporting
2. refactor version to util package
11 years ago
Chris Lu
f8b1d48f82
deal with empty learder() cases
11 years ago
Chris Lu
41143b3b78
toughen weedfs clustering, adding synchronizing max volume id among
peers in order to avoid the same volume id being assigned twice
1. moving raft.Server to topology
2. adding max volume id command for raft
11 years ago
Chris Lu
054374c765
in progress, trying to make benchmark working better to reuse http
connections.
11 years ago
Chris Lu
cd10c277b2
can now delete a collection! Is this a dangerous feature? Only enabling
deleting "benchmark" collections for now.
11 years ago
Chris Lu
7283d68e32
fixed redirection error
11 years ago
Chris Lu
27c74a7e66
Major:
change replication_type to ReplicaPlacement, hopefully cleaner code
works for 9 possible ReplicaPlacement
xyz
x : number of copies on other data centers
y : number of copies on other racks
z : number of copies on current rack
x y z each can be 0,1,2
Minor:
weed server "-mdir" default to "-dir" if empty
11 years ago
Chris Lu
ef4c2c0d1e
cleaner cluster messages
11 years ago
Chris Lu
69a7de1ad8
optimize for submission from clustered master nodes
11 years ago
Chris Lu
e7585548a4
working auto fail-over master node
11 years ago
Chris Lu
2e2f426fe2
clean up etcd backing for sequence persistent storage. It uses some OS
specific calls and is not OS-agnostic.
11 years ago
Chris Lu
c38eee73ca
refactoring to separate master and volume server, so that these servers
can be embedded into other applications
11 years ago