The SimpleProgressLog had a number of problems:
1. It polled for progress with no attempt to determine whether progress could realistically be made, so:
- as the number of pending transactions grew, the proportion of useful work dropped (as many would be unable to make progress without earlier transactions completing)
- each transaction in the chain could recover only on average 1/2 poll interval behind the last transaction to complete
2. It requested full transaction state from every replica on each attempt
3. It maintained a lot of in-memory state
4. Polling happened en-masse, allowing for little per-transaction control
We also separately maintained fairly expensive per-command listener state that negatively affected our command loading and caching.
The new DefaultProgressLog makes use of several new features: LocalListeners, RemoteListeners, Timers and Await messages.
- LocalListeners provide a memory-efficient collection for managing each CommandStore<E2><80><99>s transaction listeners, with dedicated record keeping for inter-transaction relationships.
- RemoteListeners provide a mechanism for request/response pairs that may be separated by longer than the normal Cassandra message timeout, and require minimal state on sender and recipient. This permits replicas to cheaply update their local state machine as soon as distributed information becomes available.
The DefaultProgressLog tracks each transaction with separate timers to handle per-transaction scheduling, backoff etc, and a succinct state machine. To reduce overhead correspondence is preferentially limited to a handful of replicas, and limited to the home shard where appropriate.
patch by Benedict; reviewed by Ariel Weisberg for CASSANDRA-19870
patch by Zhongxiang Zheng; reviewed by Stefan Miklosovic, Maulin Vasavada, Cheng Wang, Jordan West for CASSANDRA-11695
Co-authored-by: Stefan Miklosovic <smiklosovic@apache.org>
Co-authored-by: Sam Tunnicliffe <samt@apache.org>
Both CASSANDRA-18120 and CASSANDRA-19946 updated cassandra_latest.yaml but not the corresponding test configs
patch by Mick Semb Wever; reviewed by Brandon Williams, Tiago L. Alves for CASSANDRA-20002
This patch adds a new configuration file, cassandra_latest.yaml, which changes
some of the settings to use new features we usually recommend to users,
including:
- Big Trie-Indexed SSTables
- Trie memtables
- Unified Compaction Strategy
- Storage Attached Index
- Off-heap objects memtables
- Direct-write commit log
- No storage compatibility
These changes are applied in new test targets for unit, in-jvm and python dtests that are
run pre-commit. When changes are made to the new config file, python dtest will use it
directly, but unit and in-jvm dtests require updates, respectively, to latest_diff.yaml
and InstanceConfig.java.
patch by Branimir Lambov; reviewed by Berenguer Blasi for CASSANDRA-18753
patch by Claude Warren; reviewed by Stefan Miklosovic, Jacek Lewandowski, Michael Semb Wever for CASSANDRA-16565
Co-authored-by: Stefan Miklosovic <smiklosovic@apache.org>
Patch by Amit Pawar and Jacek Lewandowski; reviewed by Branimir Lambov and Maxwell Guo for CASSANDRA-18464
Co-authored-by: Amit Pawar <Amit.Pawar@amd.com>
Co-authored-by: Jacek Lewandowski <lewandowski.jacek@gmail.com>
An overview of the core components can be found in the included
TransactionalClusterMetadata.md
patch by Alex Petrov, Marcus Eriksson and Sam Tunnicliffe; reviewed by
Alex Petrov, Marcus Eriksson and Sam Tunnicliffe for CASSANDRA-18330
Co-authored-by: Marcus Eriksson <marcuse@apache.org>
Co-authored-by: Alex Petrov <oleksandr.petrov@gmail.com>
Co-authored-by: Sam Tunnicliffe <samt@apache.org>
The approach is documented in the included UnifiedCompactionStrategy.md.
Also included:
- Generalize prettyPrint to prettyPrintBinary (with 2^10 multiplier,
e.g. MiB) and prettyPrintDecimal (with 1000 multiplier, e.g. ms)
that cover the full range of double, and add a parseHumanReadable
method that can read both.
- CASSANDRA-18123: Fix invalid reuse of metadata collector during flushing
- Fix invalid token range splitting with non-unit weights
- Add flushSizeOnDisk metric
- Add ability to change compaction default in YAML
patch by Alex Sorokoumov, Branimir Lambov, Dimitar Dimitrov and Stefania Alborghetti
reviewed by Alex Sorokoumov, Jaroslaw Grabowski and Maxwell Guo for CASSANDRA-18397
Documentation in the included BTIFormat.md.
patch by Branimir Lambov and Jacek Lewandowski; reviewed by Caleb Rackliffe and Maxwell Guo for CASSANDRA-18398
Blake Eggleston