[doc](query-cache) Document incremental merge in cloud mode#3999
Open
asdf2014 wants to merge 1 commit into
Open
[doc](query-cache) Document incremental merge in cloud mode#3999asdf2014 wants to merge 1 commit into
asdf2014 wants to merge 1 commit into
Conversation
Follow-up to the incremental merge documentation for the cloud (compute-storage decoupled) support: - Condition 3 now states that the tablet view must be syncable to the queried version. This is trivial on local storage; on cloud the per-tablet decision first syncs the view (rowsets, and for merge-on-write tables the delete bitmap) with a fast-fail budget (query_cache_decision_sync_timeout_ms) so a slow meta service cannot hold the decision's admission thread. - A note explains that query_freshness_tolerance_ms and enable_prefer_cached_rowset make a query read a version-inexact view, so it does not use incremental merge and, on cloud, also skips the cache write-back. - Adds the query_cache_decision_sync_timeout_ms BE config row and the query_cache_decision_sync_time_ms metric, and lists the two new cloud fallback reasons (cloud rowset sync failed, cloud rowset sync timed out). - Updates the decision-flow diagram accordingly. Also corrects the JOIN caching FAQ (unrelated to the cloud work): a join in the cached subtree disables the cache structurally, independent of runtime filters, while the aggregate-first-then-join rewrite additionally needs runtime_filter_mode=OFF for every joined table to be cached, not just the build side.
16 tasks
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Versions
Languages
Docs Checklist
Documents the cloud (compute-storage decoupled) support for Query Cache incremental merge: condition 3 now covers the per-tablet view sync and its fast-fail budget (
query_cache_decision_sync_timeout_ms), adds the be.conf config row and thequery_cache_decision_sync_time_msmetric, lists the two new cloud fallback reasons, adds a note onquery_freshness_tolerance_ms/enable_prefer_cached_rowset, and updates the decision-flow diagram. Also corrects the JOIN caching FAQ (unrelated to the cloud work): a join in the cached subtree disables the cache structurally, while the aggregate-first-then-join rewrite additionally needsruntime_filter_mode=OFFfor every joined table to be cached.Code PR: apache/doris#65788