You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Dec 22, 2020. It is now read-only.
I am not seeing a new state event fired from nodes that are added to the cluster via the .join method
Scenario:
I have a 2 node skiff cluster. node1 initializes with peers set to [] and become the leader, then sets an entry in a non-skiff controlled shared database saying it is the leader
node2 starts and checks the non-skiff controlled database for the leader and starts with peers set to [leaders muliaddr], then sends an event via socket.io to node1 requesting to join and sending its multiaddr address.
node1 receives the request and initiates a .join
at this point the joined event fires on both node1 and node2 but node2 does not fire a new state event.
when taking node1 down, node2 then fires a new state indicating that it is now a candidate
is this expected behavior for the .join method? If not, what is the best method for dynamically adding peers to a cluster w/o restarting each node with a complete/updated peer list?
additionally, when using the address 127.0.0.1 for both node1 and node2 on separate processes and different ports gives the following error
Error: timedout waiting for consensus
at Timeout.onTimeout (/***/node_modules/skiff/lib/states/leader.js:122:19)
at tryOnTimeout (timers.js:224:11)
at Timer.listOnTimeout (timers.js:198:5) code: 'ETIMEOUT'
The text was updated successfully, but these errors were encountered:
Sign up for freeto subscribe to this conversation on GitHub.
Already have an account?
Sign in.
I am not seeing a
new state
event fired from nodes that are added to the cluster via the.join
methodScenario:
I have a 2 node skiff cluster.
node1
initializes with peers set to[]
and become the leader, then sets an entry in a non-skiff controlled shared database saying it is the leadernode2
starts and checks the non-skiff controlled database for the leader and starts with peers set to[leaders muliaddr]
, then sends an event via socket.io tonode1
requesting to join and sending its multiaddr address.node1
receives the request and initiates a.join
at this point the
joined
event fires on bothnode1
andnode2
butnode2
does not fire anew state
event.when taking
node1
down,node2
then fires anew state
indicating that it is now a candidateis this expected behavior for the
.join
method? If not, what is the best method for dynamically adding peers to a cluster w/o restarting each node with a complete/updated peer list?additionally, when using the address
127.0.0.1
for bothnode1
andnode2
on separate processes and different ports gives the following errorThe text was updated successfully, but these errors were encountered: