stinglet pangaea integration - #3
Draft
VijayP30 wants to merge 3 commits into
Draft
Conversation
…failure When a pod cannot fit because of insufficient zNUMA (far-memory) capacity, call out to Pangaea DCMFM Agent (POST /api/v1/memblocks) asking it to online more, then retry the combo search once with capacity refreshed from sysfs. Falls through to the existing NoNUMAComboFound behavior unchanged if DCMFM_AGENT_ADDR is unset or DCMFM has nothing to give. Previously the topology snapshot taken at kubelet startup was the ceiling for the lifetime of the process, even though CXL/far memory can grow.
…e request tryGrowFarMemCapacity was called with req.farMem -- the containers entire far-mem ask -- instead of what was actually missing. A pod that mostly fits (e.g. 94Gi requested, ~92Gi already free) was asking DCMFM to grow the full 94Gi from scratch instead of the ~2Gi shortfall, which a real or fake FAM device sized for incremental growth can never satisfy in one shot. Compute the shortfall as the ask minus everything already free across every zNUMA node before calling DCMFM.
The "dedupe a group of nNUMAs neighbor zNUMAs via a map" pattern was duplicated three times: once in findNUMACombo() (filtered to nNUMAs with free CPUs) and twice in candidateBetterThanCurrent() (unfiltered, once for the candidate group and once for the current one). Pulled into one shared helper, uniqueNeighborZNUMAs(nNUMAsGrp, onlyFreeCPUNNUMAs), with each call site keeping its existing filtering behavior via the second argument -- no behavior change, just removes the duplication. Mirrors the same extraction proposed to #4 (against the stinglet branch), applied here too so this branch stays in sync with that upstream refactor.
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.
No description provided.