--- title: Cassandra 6.0 brings ACID transactions, ClickHouse... description: A curated roundup of major database news. Apache Cassandra 6.0 alpha introduces Accord, a leaderless consensus protocol enabling multi-partition ACID... canonical: https://daily.dev/posts/cassandra-6-0-brings-acid-transactions-clickhouse-dominates-the-analytics-stories-0vv8qnbyv twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Cassandra 6.0 brings ACID transactions, ClickHouse dominates the analytics stories | daily.dev og:description: A curated roundup of major database news. Apache Cassandra 6.0 alpha introduces Accord, a leaderless consensus protocol enabling multi-partition ACID... og:url: https://daily.dev/posts/cassandra-6-0-brings-acid-transactions-clickhouse-dominates-the-analytics-stories-0vv8qnbyv og:image: https://api.daily.dev/og/posts/0vv8QnbyV.png og:image:alt: Cassandra 6.0 brings ACID transactions, ClickHouse dominates the analytics stories og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Cassandra 6.0 brings ACID transactions, ClickHouse dominates the analytics stories **[Databases Digest](https://daily.dev/sources/databases_digest)** · 5 min read · 0 upvotes · 0 comments ## Summary A curated roundup of major database news. Apache Cassandra 6.0 alpha introduces Accord, a leaderless consensus protocol enabling multi-partition ACID transactions with strict serializable isolation, alongside Transactional Cluster Metadata replacing Gossip Protocol and built-in repair scheduling. Multiple production ClickHouse writeups from Canva, Netflix, and Lyft detail massive-scale observability and analytics deployments with concrete performance numbers. A deep dive into PostgreSQL's fsyncgate history explains the data_sync_retry GUC and why it should stay off. Salesforce's Data 360 team describes scaling from 1 trillion to 120 trillion monthly rows by adopting Apache Iceberg as a common storage abstraction. Also covered: DoltLite Prolly Tree write overhead benchmarks, pg_background 2.0, Amazon RDS for Db2 12.1, Redshift snapshot billing changes, and Aurora DSQL SQLAlchemy integration requirements. ## Content **TLDR:** Apache Cassandra 6.0 alpha lands with Accord, a leaderless consensus protocol that finally delivers multi-partition ACID transactions without a coordinator — a structural shift for teams who've been working around Cassandra's consistency limits at the application layer. ClickHouse is everywhere today: Lyft, Canva, Netflix, and DeepL all published detailed migration and optimization writeups. On the PostgreSQL side, there's a deep dive into the fsyncgate-era data_sync_retry GUC that's worth reading if you run Postgres on Linux. Salesforce's Data 360 team also published how they scaled Zero Copy from 1 trillion to 120 trillion rows monthly by switching to Apache Iceberg as a common storage abstraction. --- ## Apache Cassandra 6.0 alpha ships Accord consensus, built-in repair scheduling, and TCM Cassandra 6.0 is the most operationally significant release in years. Accord is a leaderless consensus protocol delivering strict serializable isolation across multiple partitions — something teams have historically had to fake at the application layer with conditional writes and compensating transactions. Transactional Cluster Metadata replaces Gossip Protocol with a deterministic log of state transitions, which should make cluster topology changes less of a prayer. Built-in repair scheduling kills the external cron job or third-party tooling requirement that's been a persistent ops burden. There's also Zstandard dictionary compression and cursor-based compaction to reduce heap pressure. If you've been avoiding Cassandra because of its consistency model, 6.0 changes the calculus. [Read more](https://app.daily.dev/posts/GbkadaqjJ) ## Salesforce scales Zero Copy from 1 trillion to 120 trillion rows monthly via Apache Iceberg The original Query Federation model routed queries through remote warehouse compute engines — Snowflake, Databricks, BigQuery, Redshift — and that coordination overhead became the bottleneck at petabyte scale. The fix was adopting Apache Iceberg as a common storage abstraction so Data 360 operates directly against external storage rather than coordinating dual compute systems. That single architectural change removed cross-platform query execution overhead while keeping data governance intact through catalog-layer temporary credentials. The scale jump is real: under 1 trillion to 120 trillion rows processed monthly. [Read more](https://app.daily.dev/posts/QYrCoUU45) ## ClickHouse at scale: Canva cuts P90 trace latency 10x, Netflix processes 5PB daily Several detailed production writeups landed today. Canva processes 3 million spans and 3 million logs per second for 240 million MAU — migrating their observability stack to ClickHouse cut infrastructure costs by ~70% and dropped P90 trace search latency from 30 seconds to 2.5 seconds. Key moves: shifting batch processing upstream to OTel collectors, designing sort keys around actual query patterns, and using materialized view lookup tables for trace ID searches. Netflix runs at 10.6 million events per second across 5PB daily — they replaced regex-based log fingerprinting with a generated JFlex lexer for an 8-10x speedup (216μs to 23μs per fingerprint) and sharded tag maps into 31 buckets to cut filter query times from 3 seconds to under 700ms. Lyft also published their Druid-to-ClickHouse migration covering 450TB daily across batch and real-time paths. [Read more](https://app.daily.dev/feed-by-ids?id=zBwsBNvMu&id=znXrkV9Hm&id=o6yL9kVSg) ## PostgreSQL data_sync_retry GUC: the 'lose data silently' option you should leave off This is a thorough post-mortem on fsyncgate — the 2018 discovery that Linux silently discards dirty pages after writeback failures, reporting the error only once to whichever file descriptor calls fsync() first. PostgreSQL could complete a checkpoint, recycle WAL, and permanently lose data with no indication anything went wrong. The fix changed the default to PANIC on fsync() failure, because crashing and replaying from WAL is the only safe recovery path. Setting data_sync_retry = on restores the old unsafe behavior. The post also covers the Linux kernel's partial fix via errseq_t in 4.13 and the still-incomplete Direct I/O work that would fully solve the problem. If you run Postgres on Linux and haven't read about fsyncgate, this is the explainer. [Read more](https://app.daily.dev/posts/I7ZeY1kmj) --- ## Also notable - **DoltLite Prolly Tree write overhead: 4x on autocommit, 1.67x with batching:** DoltLite replaces SQLite's B-tree with a Prolly Tree for version control capabilities — the performance tax is roughly 1.20x on reads, 1.67x on batched writes, and 4x on autocommit writes, with the autocommit penalty growing as tables get larger due to per-commit root rehashing. [Read more](https://app.daily.dev/posts/AntlKgASR) - **pg_background 2.0 ships cleaner API, stricter security defaults, Postgres 19 compatibility:** pg_background 2.0 revokes PUBLIC access by default, introduces a dedicated pgbackground_role, consolidates cancellation and wait functions into single entry points, and adds PostgreSQL 19 beta compatibility — upgrade is a single ALTER EXTENSION command. [Read more](https://app.daily.dev/posts/WnRct9Qqs) - **Amazon RDS for Db2 12.1 adds AI Query Optimizer and free Community Edition capped at 4 cores/8GB:** Db2 12.1 on RDS brings an ML-based cardinality estimation optimizer and logical namespace separation for multi-tenancy; the new Community Edition is license-free but hard-capped at 4 CPU cores and 8 GB RAM, suitable for dev and small production workloads. [Read more](https://app.daily.dev/posts/7zAnvLIhS) - **Redshift manual snapshots now billed on unique blocks, not total snapshot size:** Redshift Serverless and RG instances now meter manual snapshot storage based on unique data blocks across all snapshots rather than total size per snapshot, meaning more frequent snapshots for better RPO no longer cost proportionally more. [Read more](https://app.daily.dev/posts/Z3oKeFqe9) - **Aurora DSQL with SQLAlchemy requires AUTOCOMMIT mode and application-level FK definitions:** Aurora DSQL drops ForeignKey constraint support and SAVEPOINT operations, so SQLAlchemy 2.x integrations must use AUTOCOMMIT mode, define relationships via relationship() with foreign() annotations, and use UUID primary keys — the post covers OCC retry logic with exponential backoff for write conflicts. [Read more](https://app.daily.dev/posts/HObYpiXp0) ## Similar posts on daily.dev - [For years, Apache Cassandra handed this work to your team — 6.0 takes it back](https://daily.dev/posts/for-years-apache-cassandra-handed-this-work-to-your-team-6-0-takes-it-back-gbkadaqjj) · The New Stack · 0 upvotes · 0 comments --- Tags: [#postgresql](https://daily.dev/tags/postgresql), [#clickhouse](https://daily.dev/tags/clickhouse), [#apache-cassandra](https://daily.dev/tags/apache-cassandra), [#apache-iceberg](https://daily.dev/tags/apache-iceberg) [View this post on daily.dev](https://daily.dev/posts/cassandra-6-0-brings-acid-transactions-clickhouse-dominates-the-analytics-stories-0vv8qnbyv) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/cassandra-6-0-brings-acid-transactions-clickhouse-dominates-the-analytics-stories-0vv8qnbyv","headline":"Cassandra 6.0 brings ACID transactions, ClickHouse dominates the analytics stories","text":"A curated roundup of major database news. Apache Cassandra 6.0 alpha introduces Accord, a leaderless consensus protocol enabling multi-partition ACID transactions with strict serializable isolation, alongside Transactional Cluster Metadata replacing Gossip Protocol and built-in repair scheduling. Multiple production ClickHouse writeups from Canva, Netflix, and Lyft detail massive-scale observability and analytics deployments with concrete performance numbers. A deep dive into PostgreSQL's fsyncgate history explains the data_sync_retry GUC and why it should stay off. Salesforce's Data 360 team describes scaling from 1 trillion to 120 trillion monthly rows by adopting Apache Iceberg as a common storage abstraction. Also covered: DoltLite Prolly Tree write overhead benchmarks, pg_background 2.0, Amazon RDS for Db2 12.1, Redshift snapshot billing changes, and Aurora DSQL SQLAlchemy integration requirements.","url":"https://daily.dev/posts/cassandra-6-0-brings-acid-transactions-clickhouse-dominates-the-analytics-stories-0vv8qnbyv","datePublished":"2026-06-09T04:17:56.725Z","dateModified":"2026-06-09T04:18:14.589Z","author":{"@type":"Organization","name":"Databases Digest","logo":"https://media.daily.dev/image/upload/s--ir1hdV8b--/f_auto,q_auto/v1773839400/logos/databases_digest?_a=BAMAMiiu0","url":"https://daily.dev/sources/databases_digest"},"interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/databases_digest","name":"Databases Digest"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Databases Digest","item":"https://daily.dev/sources/databases_digest"},{"@type":"ListItem","position":3,"name":"Cassandra 6.0 brings ACID transactions, ClickHouse dominates the analytics stories"}]} ```