Repository navigation
[raft] helm deployment - #1808
marcadamsge wants to merge 2 commits into
Conversation
| if err != nil { | ||
| return nil, stacktrace.Propagate(err, "failed to generate request seed") | ||
| } | ||
| metaCtx := random.NewContext(timestamp.WithRequestTimestamp(ctx, time.Now()), seed) |
There was a problem hiding this comment.
This is out of scope of the PR, but I needed it to be able to test the deployment locally, otherwise the nodes failed to start up.
@MariemBaccari How are nodes suppose to get initialized? Should I split this into another PR or just remove it?
There was a problem hiding this comment.
random was removed in #1751. Are you on the latest version of upstream's master ?
There was a problem hiding this comment.
Ah, I'll try again from the latest master without that code
There was a problem hiding this comment.
It works fine now locally in a k3s without these go changes, thanks 👍
0832844 to
176bd68
Compare
the-glu
left a comment
There was a problem hiding this comment.
This only allow a local cluster to be run, correct? Do you plan to make it "clusteriable" with others clusters in a flowup PR ?
| @@ -0,0 +1,160 @@ | |||
| {{/* Exactly one datastore backend must be enabled */}} | |||
| {{- include "dss.validate-datastore" . -}} | |||
There was a problem hiding this comment.
shouldn't this be included at a more 'global' level instead of dss-raft to perform validation in a more general context ?
| # See `core-service --help` (flags: raft_snapshot_catchup_entries, | ||
| # raft_snapshot_interval_entries, raft_tick_interval, raft_election_tick, | ||
| # raft_heartbeat_tick, raft_max_size_per_msg, raft_max_inflight_msgs). | ||
| tuning: {} |
There was a problem hiding this comment.
do we want to be generic there ? If I'm not wrong, all argument are explicit in general
Implements #1739 . Some design choices:
Each pod's Raft node ID is derived at runtime from its statefulset ordinal (parsed from the pod name) plus raft.nodeIDBase.
This PR was tested by deploying the helm chart to a local k3s cluster.