GitHub user chiradip added a comment to the discussion: Metadata layer in 
cluster.

Greg — I pushed through the tradeoffs and I think we should prefer segmented 
append-only logs + snapshot-triggered truncation over a bounded circular buffer 
for metadata.

Reason: the circular ring requires strict timing guarantees (snapshot must 
finish before wrap) and a bunch of brittle guardrails (preemptive snapshot 
triggers, throttling, overflow spill, complex crash recovery). Those are doable 
but add a lot of correctness surface area. Segmented logs tolerate follower lag 
and crashes far more naturally, simplify recovery, and are easier to 
test/operate.

If we accept slightly more IO/manifest work, we get drastically simpler 
correctness. We can still keep logical domain namespaces and modular snapshots 
(MuxStateMachine inside one VSR group) and can optimize later if real 
measurements demand it. Happy to sketch segment format + compaction/truncation 
rules if that helps.

GitHub link: 
https://github.com/apache/iggy/discussions/2346#discussioncomment-14978734

----
This is an automatically sent email for [email protected].
To unsubscribe, please send an email to: [email protected]

Reply via email to