Filter worker heartbeats by per-topology timeout#8800
Open
mwkang wants to merge 1 commit into
Open
Conversation
ReportWorkerHeartbeats dropped stale local heartbeats using only the global supervisor.worker.timeout.secs. The worker-kill authority, Slot.getHbTimeoutMs(), instead uses max(topology.worker.timeout.secs, supervisor.worker.timeout.secs). A slow-but-alive worker of a topology that raised its own timeout was therefore excluded from the report one round before Slot would treat it as dead. Compute the effective timeout per topology to mirror Slot: read the localized topology conf and use max(global, min(override, max)). The topology override is bounded by worker.max.timeout.secs -- Nimbus already clamps it at submission, re-applied here defensively against the override component only so the global floor is never reduced. When the topology conf cannot be read (the orphaned worker dirs this filter targets often outlive their conf) fall back to the global timeout. The per-topology result is cached within a reporting round. Add regression tests for a longer per-topology override, the worker.max.timeout.secs cap, and the conf-read-failure fallback.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What is the purpose of the change
ReportWorkerHeartbeatsfiltered stale heartbeats using only the globalsupervisor.worker.timeout.secs, but the kill authoritySlot.getHbTimeoutMs()usesmax(topology.worker.timeout.secs, supervisor.worker.timeout.secs). A slow-but-alive worker of a topology with a longer timeout was dropped from the report one round beforeSlotwould treat it as dead.This computes the effective timeout per topology to match
Slot:max(global, min(override, worker.max.timeout.secs)), falling back to the global timeout when the topology conf can't be read.How was this patch tested
Added unit tests for the per-topology override, the
worker.max.timeout.secscap, and the conf-read-failure fallback.storm-serverbuild and checkstyle pass.Closes #8799