442cc1d16d
The assumption that we don't need to write an entry in the invalidations
file when the hashes.pkl does not exist turned out to be a premature
optimization and also wrong.
Primarily we should recognize the creation of hashes.pkl is the first
thing that happens in a part when it lands on a new primary. The code
should be optimized toward the assumption of the most common disk state.
Also, in this case the extra stat calls to check if the hashes.pkl exists
were not only un-optimized - but introducing a race.
Consider the common case:
proc 1 | proc 2
-------------------------------|---------------------------
a) read then truncate journal |
b) do work | c) append to journal
d) apply "a" to index |
The index written at "d" may not (yet) reflect the entry writen by proc
2 at "c"; however, it's clearly in the journal so it's easy to see we're
safe.
Adding in the extra stat call for the index existence check increases
the state which can effect correctness.
proc 1 | proc 2
------------------------------|---------------------------
a) no index, truncate journal |
b) do work | b) iff index exists
| c) append to journal
d) apply (or create) index |
If step "c" doesn't happen because the index does not yet exist - the
update is clearly lost.
In our case we'd skip marking a suffix as invalid when the hashes.pkl
does not exist because we know "the next time we rehash" we'll have to
os.listdir suffixes anyway. But if another process is *currently*
rehashing (and has already done it's os.listdir) instead we've just
dropped an invalidation on the floor.
Don't do that.
Instead - write down the invalidation. The running rehash is welcome to
proceed on outdated information - as long as the next pass will grab and
hash the new suffix.
Known-Issue(s):
If the suffix already exists there's an even chance the running rehash
will hash in the very update for which we want to invalidate the suffix,
but that's ok it's idempotent.
Co-Author: Pavel Kvasnička <pavel.kvasnicka@firma.seznam.cz>
Co-Author: Alistair Coles <alistair.coles@hpe.com>
Co-Author: Kota Tsuyuzaki <tsuyuzaki.kota@lab.ntt.co.jp>
Related-Change-Id: I64cadb1a3feb4d819d545137eecfc295389794f0
Change-Id: I2b48238d9d684e831d9777a7b18f91a3cef57cd1
Closes-Bug: #1651530
|
||
|---|---|---|
| api-ref/source | ||
| bin | ||
| doc | ||
| etc | ||
| examples | ||
| install-guide/source | ||
| releasenotes | ||
| swift | ||
| test | ||
| .alltests | ||
| .coveragerc | ||
| .functests | ||
| .gitignore | ||
| .gitreview | ||
| .mailmap | ||
| .manpages | ||
| .probetests | ||
| .testr.conf | ||
| .unittests | ||
| AUTHORS | ||
| babel.cfg | ||
| bandit.yaml | ||
| bindep.txt | ||
| CHANGELOG | ||
| CONTRIBUTING.rst | ||
| LICENSE | ||
| MANIFEST.in | ||
| README.rst | ||
| requirements.txt | ||
| REVIEW_GUIDELINES.rst | ||
| setup.cfg | ||
| setup.py | ||
| test-requirements.txt | ||
| tox.ini | ||
Team and repository tags
Swift
A distributed object storage system designed to scale from a single machine to thousands of servers. Swift is optimized for multi-tenancy and high concurrency. Swift is ideal for backups, web and mobile content, and any other unstructured data that can grow without bound.
Swift provides a simple, REST-based API fully documented at http://docs.openstack.org/.
Swift was originally developed as the basis for Rackspace's Cloud Files and was open-sourced in 2010 as part of the OpenStack project. It has since grown to include contributions from many companies and has spawned a thriving ecosystem of 3rd party tools. Swift's contributors are listed in the AUTHORS file.
Docs
To build documentation install sphinx
(pip install sphinx), run
python setup.py build_sphinx, and then browse to
/doc/build/html/index.html. These docs are auto-generated after every
commit and available online at http://docs.openstack.org/developer/swift/.
For Developers
Getting Started
Swift is part of OpenStack and follows the code contribution, review, and testing processes common to all OpenStack projects.
If you would like to start contributing, check out these notes to help you get started.
The best place to get started is the "SAIO - Swift All In One". This document will walk you through setting up a development cluster of Swift in a VM. The SAIO environment is ideal for running small-scale tests against swift and trying out new features and bug fixes.
Tests
There are three types of tests included in Swift's source tree.
- Unit tests
- Functional tests
- Probe tests
Unit tests check that small sections of the code behave properly. For example, a unit test may test a single function to ensure that various input gives the expected output. This validates that the code is correct and regressions are not introduced.
Functional tests check that the client API is working as expected. These can be run against any endpoint claiming to support the Swift API (although some tests require multiple accounts with different privilege levels). These are "black box" tests that ensure that client apps written against Swift will continue to work.
Probe tests are "white box" tests that validate the internal workings of a Swift cluster. They are written to work against the "SAIO - Swift All In One" dev environment. For example, a probe test may create an object, delete one replica, and ensure that the background consistency processes find and correct the error.
You can run unit tests with .unittests, functional tests
with .functests, and probe tests with
.probetests. There is an additional .alltests
script that wraps the other three.
Code Organization
- bin/: Executable scripts that are the processes run by the deployer
- doc/: Documentation
- etc/: Sample config files
- examples/: Config snippets used in the docs
- swift/: Core code
- account/: account server
- cli/: code that backs some of the CLI tools in bin/
- common/: code shared by different modules
- middleware/: "standard", officially-supported middleware
- ring/: code implementing Swift's ring
- container/: container server
- locale/: internationalization (translation) data
- obj/: object server
- proxy/: proxy server
- test/: Unit, functional, and probe tests
Data Flow
Swift is a WSGI application and uses eventlet's WSGI server. After
the processes are running, the entry point for new requests is the
Application class in swift/proxy/server.py.
From there, a controller is chosen, and the request is processed. The
proxy may choose to forward the request to a back- end server. For
example, the entry point for requests to the object server is the
ObjectController class in
swift/obj/server.py.
For Deployers
Deployer docs are also available at http://docs.openstack.org/developer/swift/. A good starting point is at http://docs.openstack.org/developer/swift/deployment_guide.html
There is an ops runbook that gives information about how to diagnose and troubleshoot common issues when running a Swift cluster.
You can run functional tests against a swift cluster with
.functests. These functional tests require
/etc/swift/test.conf to run. A sample config file can be
found in this source tree in test/sample.conf.
For Client Apps
For client applications, official Python language bindings are provided at http://github.com/openstack/python-swiftclient.
Complete API documentation at http://docs.openstack.org/api/openstack-object-storage/1.0/content/
There is a large ecosystem of applications and libraries that support and work with OpenStack Swift. Several are listed on the associated projects page.
For more information come hang out in #openstack-swift on freenode.
Thanks,
The Swift Development Team