c8397f0287
The motivation for tablet resizing is that we want to keep the average tablet size reasonable, such that load rebalancing can remain efficient. Too large tablet makes migration inefficient, therefore slowing down the balancer.
If the avg size grows beyond the upper bound (split threshold), then balancer decides to split. Split spans all tablets of a table, due to power-of-two constraint.
Likewise, if the avg size decreases below the lower bound (merge threshold), then merge takes place in order to grow the avg size. Merge is not implemented yet, although this series lays foundation for it to be impĺemented later on.
A resize decision can be revoked if the avg size changes and the decision is no longer needed. For example, let's say table is being split and avg size drops below the target size (which is 50% of split threshold and 100% of merge one). That means after split, the avg size would drop below the merge threshold, causing a merge after split, which is wasteful, so it's better to just cancel the split.
Tablet metadata gains 2 new fields for managing this:
resize_type: resize decision type, can be either of "merge", "split", or "none".
resize_seq_number: a sequence number that works as the global identifier of the decision (monotonically increasing, increased by 1 on every new decision emitted by the coordinator).
A new RPC was implemented to pull stats from each table replica, such that load balancer can calculate the avg tablet size and know the "split status", for a given table. Avg size is aggregated carefully while taking RF of each DC into account (which might differ).
When a table is done splitting its storage, it loads (mirror) the resize_seq_number from tablet metadata into its local state (in another words, my split status is ready). If a table is split ready, coordinator will see that table's seq number is the same as the one in tablet metadata. Helps to distinguish stale decisions from the latest one (in case decisions are revoked and re-emited later on). Also, it's aggregated carefully, by taking the minimum among all replicas, so coordinator will only update topology when all replicas are ready.
When load balancer emits split decision, replicas will listen to need to split with a "split monitor" that is awakened once a table has replication metadata updated and detects the need for split (i.e. resize_type field is "split").
The split monitor will start splitting of compaction groups (using mechanism introduced here: 081f30d149) for the table. And once splitting work is completed, the table updates its local state as having completed split.
When coordinator pulls the split status of all replicas for a table via RPC, the balancer can see whether that table is ready for "finalizing" the decision, which is about updating tablet metadata to split each tablet into two. Once table replicas have their replication metadata updated with the new tablet count, they can update appropriately their set of compaction groups (that were previously split in the preparation step).
Fixes #16536.
Closes scylladb/scylladb#16580
* github.com:scylladb/scylladb:
test/topology_experimental_raft: Add tablet split test
replica: Bypass reshape on boot with tablets temporarily
replica: Fix table::compaction_group_for_sstable() for tablet streaming
test/topology_experimental_raft: Disable load balancer in test fencing
replica: Remap compaction groups when tablet split is finalized
service: Split tablet map when split request is finalized
replica: Update table split status if completed split compaction work
storage_service: Implement split monitor
topology_cordinator: Generate updates for resize decisions made by balancer
load_balancer: Introduce metrics for resize decisions
db: Make target tablet size a live-updateable config option
load_balancer: Implement resize decisions
service: Wire table_resize_plan into migration_plan
service: Introduce table_resize_plan
tablet_mutation_builder: Add set_resize_decision()
topology_coordinator: Wire load stats into load balancer
storage_service: Allow tablet split and migration to happen concurrently
topology_coordinator: Periodically retrieve table_load_stats
locator: Introduce topology::get_datacenter_nodes()
storage_service: Implement table_load_stats RPC
replica: Expose table_load_stats in table
replica: Introduce storage_group::live_disk_space_used()
locator: Introduce table_load_stats
tablets: Add resize decision metadata to tablet metadata
locator: Introduce resize_decision
ScyllaDB Documentation
This repository contains the source files for ScyllaDB Open Source documentation.
- The
devfolder contains developer-oriented documentation related to the ScyllaDB code base. It is not published and is only available via GitHub. - All other folders and files contain user-oriented documentation related to ScyllaDB Open Source and are sources for docs.scylladb.com.
To report a documentation bug or suggest an improvement, open an issue in GitHub issues for this project.
To contribute to the documentation, open a GitHub pull request.
Key Guidelines for Contributors
- Follow the ScyllaDB Style Guide.
- The user documentation is written in reStructuredText (RST) - a plaintext markup language similar to Markdown. If you're not familiar with RST, see ScyllaDB RST Examples.
- The developer documentation is written in Markdown. See Basic Markdown Syntax for reference.
Creating Knowledge Base Articles
The kb/ directory holds source files for knowledge base articles in the Knowledge Base section of the ScyllaDB documentation.
The kb/kb_common subdirectory contains a template for knowledge base articles to help you create new articles.
To create a new knowledge base article (KB):
- Copy the
kb-article-template.rstfile from/kb/kb_commonto/kband rename it with a unique name. - Open the new file and fill in the required information.
- Remove what is not needed.
- Run
make previewto build the docs and preview them locally. - Send a PR with "KB" in its title.
Building User Documentation
Prerequisites
- Python 3. Check your version with
$ python --version. - poetry 1.12 or later
- make
Mac OS X
You must have a working Homebrew in order to install the needed tools.
You also need the standard utility make.
Check if you have these two items with the following commands:
brew help
make -h
Linux Distributions
Building the user docs should work out of the box on most Linux distributions.
Windows
Use "Bash on Ubuntu on Windows" for the same tools and capabilities as on Linux distributions.
Building the Docs
- Run
make previewto build the documentation. - Preview the built documentation locally at http://127.0.0.1:5500/.
Cleanup
You can clean up all the build products and auto-installed Python stuff with:
make pristine
Information for Contributors
If you are interested in contributing to Scylla docs, please read the Scylla open source page at http://www.scylladb.com/opensource/ and complete a Scylla contributor agreement if needed. We can only accept documentation pull requests if we have a contributor agreement on file for you.
Third-party Documentation
-
Do any copying as a separate commit. Always commit an unmodified version first and then do any editing in a separate commit.
-
We already have a copy of the Apache license in our tree, so you do not need to commit a copy of the license.
-
Include the copyright header from the source file in the edited version. If you are copying an Apache Cassandra document with no copyright header, use:
This document includes material from Apache Cassandra.
Apache Cassandra is Copyright 2009-2014 The Apache Software Foundation.