Fix log formatting, heartbeat message handling, and node state management#1053
Merged
Conversation
The prefix had 2 chars for 2 nodes instead of only 1 char.
Instead of d.clusterData.Cluster.Config.Nodes, which is empty when this func is called (because ccfg has not published its first cluster.Config). This fixes this bogus startup event, used to format a bogus NodeRejoin event, used to freeze the node when peer node was frozen when we were stopped: at: "2026-06-22T18:42:57.019766924+02:00" data: installed_gens: dev2n1: 16 dev2n2: 11 dev2n3: 1 joined_nodes: - dev2n3 - dev2n1 labels: node: dev2n3 new_type: patch node: dev2n3 nodes: - dev2n1 - dev2n2 - dev2n3 old_type: full id: 2 kind: HeartbeatMessageTypeUpdated
on nmon.Manager Start(), after the event subscriptions are accepted. Without this init, the Manager.clusterConfig can be a zero value.
Early in the daemon startup sequence, d.clusterData.Cluster.Config.Nodes is still a nil, but daemondata.GetHbMessageType() is already called to forge a HeartbeatMessageTypeUpdated event, then used by nmon to decide to end the rejoin period immediately if all nodes already rejoined. Nil Nodes was interpreted always has "no missing nodes".
queueFreeze() error were logged at info level.
* Also daemondata now initializes node.Monitor.State to rejoin * Publish the NodeRejoin event from the immediate rejoin codepath, so the merge peer frozen is not bypassed in this shortcut. * Log "we are late to the party, immediate rejoin" when this shortcut is entered. * Use direct <var>/node/frozen manipulation from nmon, instead of a mix of direct and callout * Better and more symmetric log entries format in rejoin period expire and peer-frozen-when-dead-merge * Log all peer frozen since last shutdown instead of just any one of those.
So a node crashing has its last_shutdown at most 10s in before the event, and peer frozen after this barrier are mirrored when the daemon starts up again.
These allow tracking the blackout begin-end time of nmon, that will be used by imon to decide if a peer instance freeze should be mirrored locally.
Fixing:
--- FAIL: TestDaemonData/Ensure_ClusterNodeData_result_is_a_deep_copy (0.00s)
main_test.go:140:
Error Trace: /home/runner/work/om3/om3/daemon/daemondata/main_test.go:140
Error: Not equal:
expected: 0
actual : 2
Test: TestDaemonData/Ensure_ClusterNodeData_result_is_a_deep_copy
main_test.go:142:
Error Trace: /home/runner/work/om3/om3/daemon/daemondata/main_test.go:142
Error: Should be false
Test: TestDaemonData
And:
--- FAIL: TestDaemonData/Ensure_node.MonitorData.GetByNode_result_is_a_deep_copy (0.00s)
main_test.go:157:
Error Trace: /root/dev/om3/daemon/daemondata/main_test.go:157
Error: Not equal:
expected: 0
actual : 2
Test: TestDaemonData/Ensure_node.MonitorData.GetByNode_result_is_a_deep_copy
Messages: State changed !
main_test.go:163:
Error Trace: /root/dev/om3/daemon/daemondata/main_test.go:163
Error: Should be false
Test: TestDaemonData
The commit adding Levelf as a backend for Infof & co added one more frame to skip. Without this patch the caller information of log entries points to the Levelf call, not useful.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.