flashbox-l1: trim lighthouse resource footprint via flags#162
Draft
MoeMahhouk wants to merge 1 commit into
Draft
flashbox-l1: trim lighthouse resource footprint via flags#162MoeMahhouk wants to merge 1 commit into
MoeMahhouk wants to merge 1 commit into
Conversation
The L1 beacon node only needs to drive the searcher's EL via the Engine API (forkchoice). It has no attached validators, serves no REST consumers, proposes no blocks, and does not need to serve historical state/blobs or light-client data to peers. Tune flags to that reality. Add to lighthouse.service ExecStart: - --target-peers 50: default is 200; the node only follows head, so cut gossip-validation CPU, bandwidth, and per-connection memory while keeping a healthy head-following margin. - --disable-light-client-server: LC server is on by default; we don't serve light clients. - --disable-upnp: no UPnP gateway exists in the TEE/cloud environment. - --hdiff-buffer-cache-size 1: default 16; each buffer ~ one BeaconState and is irrelevant for a pruning node, so cap the RAM ceiling. - --epochs-per-migration 4: default 1; batch hot->cold migration to reduce disk writes. - --prune-payloads true / --prune-blobs true: already the v8.1.3 defaults, set explicitly to document the minimal-footprint intent. Left --beacon-processor-max-workers at default so it auto-scales to the CVM's vCPU budget (a deploy-time decision, not baked into the image). Verified against a locally-built lighthouse v8.1.3 (the version pinned in mkosi.build): the exact full ExecStart is accepted by clap and boots into runtime. End-to-end Engine API drive was not exercised.
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.
The L1 beacon node only needs to drive the searcher's EL via the Engine API (forkchoice). It has no attached validators, serves no REST consumers, proposes no blocks, and does not need to serve historical state/blobs or light-client data to peers. Tune flags to that reality.
Add to lighthouse.service ExecStart:
Left --beacon-processor-max-workers at default so it auto-scales to the CVM's vCPU budget (a deploy-time decision, not baked into the image).
Verified against a locally-built lighthouse v8.1.3 (the version pinned in mkosi.build): the exact full ExecStart is accepted by clap and boots into runtime. End-to-end Engine API drive was not exercised.