Repository navigation
feat: optimize manifest writing on commit - #137
Merged
Merged
Conversation
A commit of several steps — a large transaction, or a materializer catching up — chained fast appends, one step after another: each wrote its manifests, then its manifest list, then read that list back for the next step. Four round trips to object storage a step, while committing was most of what the materializer did. Now each step's snapshot is planned in memory (the fork's `FastAppendAction::stage`), then every step's manifests are written at once, then every step's manifest list, each naming its parent's manifests and its own: a read of the table's manifest list and two rounds of writes, however many steps. V3 tables, whose snapshots take their row range from the written list, still commit step by step. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
hasyimibhar
force-pushed
the
perf/staged-commit
branch
from
October 19, 2026 13:48
d072038 to
558975a
Compare
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
A commit of several steps chained fast appends one step after another: each wrote its manifests, then its manifest list, then read that list back for the next step. Four round trips to object storage per step, and committing was 43% of the materializer's time.
Now each step's snapshot is planned in memory (polynya-dev/iceberg-rust#5,
FastAppendAction::stage;polynya-patches-v0.10.1at 0ded76c), then every step's manifests are written at once, then every step's manifest list, each naming its parent's manifests and its own: one read of the table's manifest list and two rounds of writes, however many steps.TPROC-C with cloud latency (
benchmark/tpcc,LATENCY=1),flush_rows: 100000, pinned to 2 cores, similar source load (~90k changes/s):commit_snapshots)