| Overview |  |
action() = leave | remove | {replace, node()} | {force_replace, node()}
handoff() = {atom(), integer(), atom()}
handoffs() = [handoff()]
home() = primary | fallback | undefined
index() = integer()
member() = #member_info_v2{node = atom(), status = status(), reachable = boolean(), vnodes = vnodes(), handoffs = handoffs(), ring_pct = float(), pending_pct = float(), mem_total = integer(), mem_used = integer(), mem_erlang = integer(), action = action(), replacement = node()}
members() = [member()]
n_vals() = [pos_integer()]
normalized_action() = leave | remove | replace | force_replace | stop | down
owner() = atom()
partition() = #partition_info{index = index(), partition = integer(), owner = owner(), vnodes = services(), handoffs = handoffs()}
partitions() = [partition()]
ring() = riak_core_ring:riak_core_ring()
service() = {atom(), home()}
services() = [service()]
stage_error() = nodedown | already_leaving | not_member | only_member | is_claimant | invalid_replacement | already_replacement | not_reachable | not_single_node | self_join
status() = valid | invalid | down | leaving | incompatible | transitioning
version() = integer()
vnode() = {{atom(), atom()}, atom()}
vnodes() = [vnode()]
| clear_plan/0 | Clear the staged cluster plan. |
| code_change/3 | |
| commit_plan/0 | Commit a staged cluster plan. |
| force_update/0 | Force ring/membership update. |
| get_default_n_val/0 | Return list of available n_vals. |
| get_my_info/0 | Return current nodes information. |
| get_n_vals/0 | Return list of available n_vals. |
| get_nodes/0 | Return list of cluster members. |
| get_partitions/0 | Return partition list. |
| get_plan/0 | Get the staged cluster plan. |
| get_ring/0 | Return ring. |
| get_services/0 | Return service list. |
| get_status/0 | Get overall cluster status. |
| get_version/0 | Return version. |
| handle_call/3 | |
| handle_cast/2 | Whenever there's a change to the ring, we try and update our localized view of "the world". |
| handle_info/2 | |
| init/1 | Initialize and retrieve initial state. |
| stage_change/3 | Stage a change to the cluster. |
| start_link/0 | Start service from supervisor. |
| terminate/2 |
clear_plan() -> ok | error
Clear the staged cluster plan.
code_change(Old, State, Extra) -> any()
commit_plan() -> ok | error
Commit a staged cluster plan.
force_update() -> ok
Force ring/membership update.
get_default_n_val() -> {ok, version(), pos_integer()}
Return list of available n_vals.
get_my_info() -> member()
Return current nodes information.
Return list of available n_vals.
Return list of cluster members.
get_partitions() -> {ok, version(), partitions()}
Return partition list.
get_plan() -> {ok, list(), list()} | {error, atom()}
Get the staged cluster plan.
Return ring.
get_services() -> {ok, version(), services()}
Return service list.
Get overall cluster status.
get_version() -> version()
Return version.
handle_call(X1, From, State) -> any()
handle_cast(X1, State) -> any()
Whenever there's a change to the ring, we try and update our localized view of "the world". However, ring updates can happen very often, and large rings change *a lot*. For this reason, we only want to update the ring every so often. So once we've updated the ring, a tick flag is set indicating that we can't update the ring any more until it's cleared.
handle_info(X1, State) -> any()
init(X1) -> any()
Initialize and retrieve initial state. Also, register watchers for ring and membership changes.
stage_change(Node::node(), Action::normalized_action(), Replacement::node()) -> ok | {error, stage_error()} | {badrpc, nodedown}
Stage a change to the cluster.
start_link() -> {ok, pid()}
Start service from supervisor.
terminate(Reason, State) -> any()
Generated by EDoc