31be352cb2
The benchmark first writes 1000 key/values (actually, the same key is re-used) in LMDB and read it 1000 times. Then it does the same thing using concurrency, on a 16 cores machine, with the following configurations: - 32 threads (29 readers and 3 writer) - 8 threads (6 readers and 2 writer) - 8 threads (8 readers and 0 writer) See [3] for LMDB internal locking result and [4] for rwlock result. What the result seems to say is that LMDB internal locking mechanism is doing way better (100x) when there are lot of concurrency involved. I'm kinda surprised by such factor and tried various things in case I've done something wrong, but eventually always end up with similar difference. The only explanation I have is that LMDB actually does uses an userspace RCU mechanism internally, so readers are never blocked. Using rwlock would block the readers while a writer lock is alive. LMDB doc seems to say [1] that if we use our own synchronization mechanism, we should only have one writer at the time, and hold a write lock from the beginning of the transaction up to the commit/revert (I tried holding a read lock on a write transaction then upgrade to a write lock only at commit time... but of course I end up in lot of odd crashes). I think this explain the massive congestion when using 32 threads, as lot of readers are waiting. When there is no thread congestion (the 2 later configuration with 8 threads) using the rwlock actually seems a little bit faster (same order of magnitude though). I guess this might be explained by the fact that, while LMDB doesn't lock readers, there is still a quick lock-unlock to maintains a reader table (and eventually free unused old versions once a writer commit). [2] [1] http://www.lmdb.tech/doc/group__mdb.html#:~:text=must%20ensure%20that%20no%20readers%20are%20using%20old%20transactions%20while%20a%20writer%20is%20active [2] http://www.lmdb.tech/doc/group__readers.html [3] two runs using LMDB internal locking (HEAD~2) sequential writes and read avg write 0.005878 ms avg read 0.001640 ms concurrent 29 readers and 3 writers avg write 0.746217 ms avg read 6.644507 ms concurrent 6 readers and 2 writers avg write 0.077576 ms avg read 0.177316 ms concurrent 8 readers and 0 writer avg write 0.000000 ms avg read 0.273563 ms sequential writes and read avg write 0.005799 ms avg read 0.001656 ms concurrent 29 readers and 3 writers avg write 0.737782 ms avg read 6.616162 ms concurrent 6 readers and 2 writers avg write 0.075893 ms avg read 0.178631 ms concurrent 8 readers and 0 writer avg write 0.000000 ms avg read 0.303209 ms [4] two runs using rwlock (HEAD~1) sequential writes and read avg write 0.005931 ms avg read 0.001658 ms concurrent 29 readers and 3 writers avg write 17.249417 ms avg read 163.805094 ms concurrent 6 readers and 2 writers avg write 0.040559 ms avg read 0.168439 ms concurrent 8 readers and 0 writer avg write 0.000000 ms avg read 0.200699 ms sequential writes and read avg write 0.007423 ms avg read 0.001656 ms concurrent 29 readers and 3 writers avg write 20.312584 ms avg read 190.565366 ms concurrent 6 readers and 2 writers avg write 0.026039 ms avg read 0.147365 ms concurrent 8 readers and 0 writer avg write 0.000000 ms avg read 0.174471 ms
BIND 9
Contents
- Introduction
- Reporting bugs and getting help
- Contributing to BIND
- Building BIND
- Automated testing
- Documentation
- Acknowledgments
Introduction
BIND (Berkeley Internet Name Domain) is a complete, highly portable implementation of the Domain Name System (DNS) protocol.
The BIND name server, named, can act as an authoritative name
server, recursive resolver, DNS forwarder, or all three simultaneously. It
implements views for split-horizon DNS, automatic DNSSEC zone signing and
key management, catalog zones to facilitate provisioning of zone data
throughout a name server constellation, response policy zones (RPZ) to
protect clients from malicious data, response rate limiting (RRL) and
recursive query limits to reduce distributed denial of service attacks,
and many other advanced DNS features. BIND also includes a suite of
administrative tools, including the dig and delv DNS lookup tools,
nsupdate for dynamic DNS zone updates, rndc for remote name server
administration, and more.
BIND 9 began as a complete rewrite of the BIND architecture that was used in versions 4 and 8. Internet Systems Consortium (https://www.isc.org), a 501(c)(3) US public benefit corporation dedicated to providing software and services in support of the Internet infrastructure, developed BIND 9 and is responsible for its ongoing maintenance and improvement. BIND is open source software licensed under the terms of the Mozilla Public License, version 2.0.
For a detailed list of changes made throughout the history of BIND 9, see the changelog.
For up-to-date versions and release notes, see https://www.isc.org/download/.
For information about supported platforms, see the "Supported Platforms" section in the BIND 9 Administrator Reference Manual.
Reporting bugs and getting help
To report non-security-sensitive bugs or request new features, you may open an issue in the BIND 9 project on the ISC GitLab server at https://gitlab.isc.org/isc-projects/bind9.
Please note that, unless you explicitly mark the newly created issue as
"confidential," it will be publicly readable. Please do not include any
information in bug reports that you consider to be confidential unless
the issue has been marked as such. In particular, if submitting the
contents of your configuration file in a non-confidential issue, it is
advisable to obscure key secrets; this can be done automatically by
using named-checkconf -px.
For information about ISC's Security Vulnerability Disclosure Policy and
information about reporting potential security issues, please see
SECURITY.md.
Professional support and training for BIND are available from ISC. Contact us at https://www.isc.org/contact for more information.
To join the BIND Users mailing list, or view the archives, visit https://lists.isc.org/mailman/listinfo/bind-users.
If you're planning on making changes to the BIND 9 source code, you may also want to join the BIND Workers mailing list, at https://lists.isc.org/mailman/listinfo/bind-workers.
Contributing to BIND
ISC maintains a public git repository for BIND; details can be found at https://www.isc.org/sourceaccess/.
Information for BIND contributors can be found in the following files:
- General information: CONTRIBUTING.md
- Code of Conduct: CODE_OF_CONDUCT.md
- BIND 9 code style: doc/dev/style.md
- BIND architecture and developer guide: doc/dev/dev.md
Patches for BIND may be submitted as merge requests on the ISC GitLab server.
By default, external contributors do not have the ability to fork BIND on the GitLab server; if you wish to contribute code to BIND, you may request permission to do so. Thereafter, you can create git branches and directly submit requests that they be reviewed and merged.
If you prefer, you may also submit code by opening a
GitLab issue and
including your patch as an attachment, preferably generated by
git format-patch.
Building BIND 9
For information about building BIND 9, see the "Building BIND 9" section in the BIND 9 Administrator Reference Manual.
Automated testing
A system test suite can be run with make check. The system tests require
you to configure a set of virtual IP addresses on your system (this allows
multiple servers to run locally and communicate with each other). These
IP addresses can be configured by running the command
bin/tests/system/ifconfig.sh up as root.
Some tests require Perl and the Net::DNS and/or IO::Socket::IP modules,
and are skipped if these are not available. Some tests require Python
and the dnspython module and are skipped if these are not available.
See bin/tests/system/README for further details.
Unit tests are implemented using the CMocka unit testing framework. To build
them, use configure --with-cmocka. Execution of tests is done by the automake
parallel test driver; unit tests are also run by make check.
Documentation
The BIND 9 Administrator Reference Manual (ARM) is included with the source
distribution, and in .rst format, in the doc/arm
directory. The HTML version is automatically generated and can
be viewed at https://bind9.readthedocs.io/en/latest/index.html.
The PDF version can be built by running:
cd doc/arm/
sphinx-build -b latex . pdf/
make -C pdf/ all-pdf
The above requires TeX Live in order to work. The PDF will be written to
doc/arm/pdf/Bv9ARM.pdf.
Man pages for some of the programs in the BIND 9 distribution are also included in the BIND ARM.
Frequently (and not-so-frequently) asked questions and their answers can be found in the ISC Knowledgebase at https://kb.isc.org.
Additional information on various subjects can be found in other
README files throughout the source tree.
Bug report identifiers
Most notes in the ARM Changelog appendix include a reference to a bug report or
issue number. Prior to 2018, these were usually of the form [RT #NNN]
and referred to entries in the "bind9-bugs" RT database, which was not open
to the public. More recent entries use the form [GL #NNN] or, less often,
[GL !NNN], which, respectively, refer to issues or merge requests in the
GitLab database. Most of these are publicly readable, unless they include
information which is confidential or security-sensitive.
To look up a GitLab issue by its number, use the URL https://gitlab.isc.org/isc-projects/bind9/issues/NNN. To look up a merge request, use https://gitlab.isc.org/isc-projects/bind9/merge_requests/NNN.
In rare cases, an issue or merge request number may be followed with the letter "P". This indicates that the information is in the private ISC GitLab instance, which is not visible to the public.
Acknowledgments
-
The original development of BIND 9 was underwritten by the following organizations:
Sun Microsystems, Inc. Hewlett Packard Compaq Computer Corporation IBM Process Software Corporation Silicon Graphics, Inc. Network Associates, Inc. U.S. Defense Information Systems Agency USENIX Association Stichting NLnet - NLnet Foundation Nominum, Inc. -
This product includes software developed by the OpenSSL Project for use in the OpenSSL Toolkit. https://www.OpenSSL.org/
-
This product includes cryptographic software written by Eric Young (eay@cryptsoft.com).
-
This product includes software written by Tim Hudson (tjh@cryptsoft.com).