Author: zznate
Date: Wed Oct 17 00:25:53 2018
New Revision: 1844054
URL: http://svn.apache.org/viewvc?rev=1844054&view=rev
Log:
CASSANDRA-14631 (update) - add missing feed.xml
Added:
cassandra/site/publish/feed.xml
Added: cassandra/site/publish/feed.xml
URL:
http://svn.apache.org/viewvc/cassandra/site/publish/feed.xml?rev=1844054&view=auto
==============================================================================
--- cassandra/site/publish/feed.xml (added)
+++ cassandra/site/publish/feed.xml Wed Oct 17 00:25:53 2018
@@ -0,0 +1,115 @@
+<?xml version="1.0" encoding="utf-8"?><feed
xmlns="http://www.w3.org/2005/Atom" ><generator uri="https://jekyllrb.com/"
version="3.4.3">Jekyll</generator><link
href="http://cassandra.apache.org/feed.xml" rel="self"
type="application/atom+xml" /><link href="http://cassandra.apache.org/"
rel="alternate" type="text/html"
/><updated>2018-10-17T13:22:37+13:00</updated><id>http://cassandra.apache.org/</id><title
type="html">Apache Cassandra Website</title><subtitle>The Apache Cassandra
database is the right choice when you need scalability and high availability
without compromising performance. Linear scalability and proven fault-tolerance
on commodity hardware or cloud infrastructure make it the perfect platform for
mission-critical data. Cassandra's support for replicating across multiple
datacenters is best-in-class, providing lower latency for your users and the
peace of mind of knowing that you can survive regional outages.
+</subtitle><entry><title type="html">Testing Apache Cassandra 4.0</title><link
href="http://cassandra.apache.org/blog/2018/08/21/testing_apache_cassandra.html"
rel="alternate" type="text/html" title="Testing Apache Cassandra 4.0"
/><published>2018-08-21T15:00:00+12:00</published><updated>2018-08-21T15:00:00+12:00</updated><id>http://cassandra.apache.org/blog/2018/08/21/testing_apache_cassandra</id><content
type="html"
xml:base="http://cassandra.apache.org/blog/2018/08/21/testing_apache_cassandra.html"><p>With
the goal of ensuring reliability and stability in Apache Cassandra 4.0, the
projectâs committers have voted to freeze new features on September 1 to
concentrate on testing and validation before cutting a stable beta. Towards
that goal, the community is investing in methodologies that can be performed at
scale to exercise edge cases in the largest Cassandra clusters. The result, we
hope, is to make Apache Cassandra 4.0 the best-tested and most reliable major
release r
ight out of the gate.</p>
+
+<p>In the interests of communication (and hopefully more participation),
hereâs a look at some of the approaches being used to test Apache Cassandra
4.0:</p>
+
+<hr />
+
+<h4 id="replay-testing">Replay Testing</h4>
+<h5 id="workload-recording-log-replay-and-comparison">Workload
Recording, Log Replay, and Comparison</h5>
+
+<p>Replay testing allows for side-by-side comparison of a workload using
two versions of the same database. It is a black-box technique that answers the
question, âdid anything change that we didnât expect?â</p>
+
+<p>Replay testing is simple in concept: record a workload, then re-issue
it against two clusters â one running a stable release and the second running
a candidate build. Replay testing a stateful distributed system is more
challenging. For a subset of workloads, we can achieve determinism in testing
by grouping writes by CQL partition and ordering them via client-supplied
timestamps. This also allows us to achieve parallelism, as recorded workloads
can be distributed by partition across an arbitrarily-large fleet of writers.
Though linearizing updates within a partition and comparing differences does
not allow for validation of all possible workloads (e.g., CAS queries), this
subset is very useful.</p>
+
+<p>The suite of Full Query Logging (âFQLâ) tools in Apache Cassandra
enable workload recording. <a
href="https://issues.apache.org/jira/browse/CASSANDRA-14618">CASSANDRA-14618</a>
and <a
href="https://issues.apache.org/jira/browse/CASSANDRA-14619">CASSANDRA-14619</a>
will add fqltool replay and fqltool compare, enabling log replay and
comparison. Standard tools in the Apache ecosystem such as <a
href="https://spark.apache.org">Apache Spark</a> and <a
href="https://mesos.apache.org">Apache Mesos</a> can also
make parallelizing replay and comparison across large clusters of machines
straightforward.</p>
+
+<hr />
+
+<h4 id="fuzz-testing-and-property-based-testing">Fuzz Testing
and Property-Based Testing</h4>
+<h5 id="dynamic-test-generation-and-fuzzing">Dynamic Test
Generation and Fuzzing</h5>
+
+<p>Fuzz testing dynamically generates input to be passed through a
function for validation. We can make fuzz testing smarter in stateful systems
like Apache Cassandra to assert that persisted data conforms to the
databaseâs contracts: acknowledged writes are not lost, deleted data is not
resurrected, and consistency levels are respected. Fuzz testing of storage
systems to validate these properties requires maintaining a record of responses
received from the system; the development of a model representing valid legal
states of data within the database; and a validation pass to assert that
responses reflect valid states according to that model.</p>
+
+<p>Property-based testing combines fuzz testing and assertions to
explore a state space using randomly-generated input. These tests provide
dynamic input to the system and assert that its fundamental properties are not
violated. These properties can range from generic (e.g., âI can write data
and read it backâ) to specific (ârange tombstone bounds synthesized during
short-read-protection reads are properly closedâ); and from local to
distributed (e.g., âreplacing every single node in a cluster results in an
identical databaseâ). To simplify debugging, property-based testing libraries
like <a
href="https://github.com/ncredinburgh/QuickTheories">QuickTheories</a>
also provide a âshrinker,â which attempts to generate the simplest
possible failing case after detecting input or a sequence of actions that
triggers a failure.</p>
+
+<p>Unlike model checkers, property-based tests donât exhaust the state
space â but explore it until a threshold of examples is reached. This allows
for the computation to be distributed across many machines to gain confidence
in code and infrastructure that scales with the amount of computation applied
to test it.</p>
+
+<hr />
+
+<h4
id="distributed-tests-and-fault-injection-testing">Distributed
Tests and Fault-Injection Testing</h4>
+<h5 id="validating-behavior-under-fault-scenarios">Validating
Behavior Under Fault Scenarios</h5>
+
+<p>All of the above techniques can be combined with fault injection
testing to validate that the system maintains availability where expected in
fault scenarios, that fundamental properties hold, and that reads and writes
conform to the systemâs contracts. By asserting series of invariants under
fault scenarios using different techniques, we gain the ability to exercise
edge cases in the system that may reveal unexpected failures in extreme
scenarios. Injected faults can take many forms â network partitions, process
pauses, disk failures, and more.</p>
+
+<hr />
+
+<h4 id="upgrade-testing">Upgrade Testing</h4>
+<h5 id="ensuring-a-safe-upgrade-path">Ensuring a Safe Upgrade
Path</h5>
+
+<p>Finally, itâs not enough to test one version of the database.
Upgrade testing allows us to validate the upgrade path between major versions,
ensuring that a rolling upgrade can be completed successfully, and that
contents of the resulting upgraded database is identical to the original. To
perform upgrade tests, we begin by snapshotting a cluster and cloning it twice,
resulting in two identical clusters. One of the clusters is then upgraded.
Finally, we perform a row-by-row scan and comparison of all data in each
partition to assert that all rows read are identical, logging any deltas for
investigation. Like fault injection tests, upgrade tests can also be thought of
as an operational scenario all other types of tests can be parameterized
against.</p>
+
+<hr />
+
+<h4 id="wrapping-up">Wrapping Up</h4>
+
+<p>The Apache Cassandra developer community is working hard to deliver
Cassandra 4.0 as the most stable major release to date, bringing a variety of
methodologies to bear on the problem. We invite you to join us in the effort,
deploying these techniques within your infrastructure and testing the release
on your workloads. Learn more about how to get involved <a
href="http://cassandra.apache.org/community/">here</a>.</p>
+
+<p>The more that join, the better the release weâll ship
together.</p></content><author><name>the Apache Cassandra
Community</name></author><summary type="html">With the goal of ensuring
reliability and stability in Apache Cassandra 4.0, the projectâs committers
have voted to freeze new features on September 1 to concentrate on testing and
validation before cutting a stable beta. Towards that goal, the community is
investing in methodologies that can be performed at scale to exercise edge
cases in the largest Cassandra clusters. The result, we hope, is to make Apache
Cassandra 4.0 the best-tested and most reliable major release right out of the
gate.</summary></entry><entry><title type="html">Hardware-bound Zero Copy
Streaming in Apache Cassandra 4.0</title><link
href="http://cassandra.apache.org/blog/2018/08/07/faster_streaming_in_cassandra.html"
rel="alternate" type="text/html" title="Hardware-bound Zero Copy Streaming in
Apache Cassandra 4.0" /><published>20
18-08-07T07:00:00+12:00</published><updated>2018-08-07T07:00:00+12:00</updated><id>http://cassandra.apache.org/blog/2018/08/07/faster_streaming_in_cassandra</id><content
type="html"
xml:base="http://cassandra.apache.org/blog/2018/08/07/faster_streaming_in_cassandra.html"><p>Streaming
in Apache Cassandra powers host replacement, range movements, and cluster
expansions. Streaming plays a crucial role in the cluster and as such its
performance is key to not only the speed of the operations its used in but the
clusterâs health generally. In Apache Cassandra 4.0, we have introduced an
improved streaming implementation that reduces GC pressure and increases
throughput several folds and are now limited, in some cases, only by the disk /
network IO (See: <a
href="https://issues.apache.org/jira/browse/CASSANDRA-14556">CASSANDRA-14556</a>).</p>
+
+<p><img
src="data:image/png;base64,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
SVkhTo7yUPFPjOvFWCj1mWVmp4YWUSYNxinKedeEn31NZeCI7XFOUluk8IeALs2BQ/yAYMO4IBJRXZ456dgtb+NsfYVmN/kY4REN2DDdTqOP2biqwNMKKJ8Ops/rvo6lvs7Iiia69r1lCWifJMyEwtm8lpMscDl1mpB5z6aZxgxXmGMP0D3/h8jhOpUUt/b7Qew214YfgOrOu+rUzykCsJGHylK+ELzZcGYwhA3bI627hkpgvQzHCQRqrX4H67qS4AfiWQ/fEgGvbWuSw3baXhiklJUVqUVGpjkEhl2T+uulnnmpWNUIPG7/BozkeUia4MLgdrPqkUzFWXz9yzYeR3fFu8jq8I72hwSuezz5UF9rIHI6zU8G304G70mEJxwGI2ykgxunwm1WZHixnhvvsyjlX7LQizqdA0yPvGLkdCwJSAokFrSTkFUNz0mUKfea9W8b3V8PKTyr1yl9kZLsG11UHQiDM/w3UBp+oUis3ekdupmWBeSJDUdm8QmPuXK+hFA/nD295gPWD/eJMzMkmM7xth+zUWLJi4eXA9ukkICusOzgJrbZ3IjUIrcysavmKFjoPjd3JTL9ay9vB3s2XrYsPAUme8eGW8FKCNI8amr8y/kIEMJ/KQqw4PdvpbKgJftZQHZzjchhPXH/xmq9FI4TxyJRSTYs7hvpQiEL3BsWoEgqcozhcKgxFJMbZdrRBFiySNGs9TnZ6uouFdINtrW5gedhvaxHOxJSChTgQsu7baO6T+MjUyRGQ5NnBF5gMPsOHb23UZ9LQvEqvGotm3eDO0C7y18PAw5Q1kDb/nJHTY9PUcx7igpgg22IM60vh02l6ch7Tma2lnCh1QzkXUhEOciu7LV2UiDhd+FSAMF8dcyYr7FOA1hrs4dfL2O3vfMn6Z7hYDX7bNlnSJ5bqGMpoXajikSE5dLft1UpIwyR5JgTG2CsRpFlc7OVDu2WbJuTCzEBD8+mZue6T6ir91b760EJD0f8yc9S6
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
vq5Uw9xHWhpaavAt3Jd5GGJQGIRoCHp9d/SaW46XJHWgTC+08WJk/A1hUeakYSJAJ5RPwj7XnnqVRCnyjZvbhX/1qVGkqYQFCDvnrXnI1rK46gYxElh6bqsxBkJKobw/hBzufsGnUZpP8Qx5EN4ulFlkgjYBQGSpKxZNQHd91cQ5wWWZ0xXlTgjrwy3w5OXAYIU35WzeN0vMXrUIQS1+gy3nIErUL2hXiue747JlbCqu9MwTcv03I88fdfedtAbHEak0+sCgQc5FHSj0g0rk0TADghQnEuk3IXr5zJnWpHhr0No25R1fG8PRIEPdKCgNhiB78/+w7pvR2PHaJ48I95WmK1IMw4GwzJlKpbbo/mpXSckUFgrXZ5iKJ9n8a5YN2xqd0u2PuURIJUbpk3nLXr2Kth6vZA40SU+N1C+3I+6uBhJw2sIEwTS1KDxaM7CZ/PMUWTzBqTmyZNb1hnLuXvtzfDjnIw56rTGUHQ60qMa1UV+QA9Mor+hGHd1X7T2Ih7Rmm5cmSQCHYUAjRyhcstf/Hw/+L3/iUdmN6dbSuJs+ppoJCBCrXEqmHQ5z2JOW20Sr6bJkx56DD0LIL7iHbWAlsIwjURyKNo05nwvBUygN5czyNQHKLgCf3PJ4XsLkMlD7YoAIk3w+o3gMkxNHIA4unLk2BrgEBAhKOrM5bo2d/Ha63h2r7dJIagJ8gRBWlGjESl+AVjYwQPkde5ZB61BGu1xzD7yh2GBPzVYXf17XkgO36PFTuZLJAIkAMHdJn/husm66rja8NWRj0yTJJDI06Z8XVzY4TEtSGD0Zt29qoC7Hzbh/3k8edISokgFi9b/FH6Ml0JHAtDlcD2Gm0IFgSK7MjNv0ZpT+fBdrNQZQyUyq0QgbgSIAMhDZtHLOZjQDQUcRp1YnY2PHuOutEsVxPA9ACEobaDD4TB9P7Ge0LHpaPK0QO+24JneYaZ7MQxFfvyRQ89jcWvpN1wfDOhNPFmAbT7PyFf0ky
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
dQDIBKEmSB6sEZRGjB2cp2AWz9rHySNpPGrbGIZVGOwPNuLU/CAci4LgAHPvmTlqcseyZEnyUfOYLDGZmmCpBoNw+lUD7a9FDT6Ax0doMmo0Lh973MeOXpfMmvAVuk9++ZMnQhISEmtmzZ9e773uwNroeorUqpjspgIFew1p3AscGoKjxJd11qvneMl/idzl2kfhfuGC7iEUjgE/BhR4cYrwrpgsLg6hpQVQPgvG7TdDaVWF0hO6dAzgACCOuCyB66LI1I8OO8aCTkDQWzpT7P/FpIoxceekw6QCVIKY/7ASicyqum1KB9zXwiy5atOhVDBqDQfjJwWDwQxrVERSAdqQlqme6lQJonxp+9u233z7MNM3T0VbX4xoo6Q0Oz1j6zHxooRYxH1YUmiGALFRy/ZkRInUiPOtCTOfgNnfqjnPFPqy4Em9NEjQxgh0M7RfbvQURUOCDkCX+13OnfnBoKHS+Ha5fQqMWRFwatVCxfrkiiQjOhQmDFS3llvmL8nkTpwvgxEjl5OXJkfvBxMTE7wAwHyOyEHDm5OTo1FC9ZFLnigK9SQEa1Ak48TspEolsx3UB2u1QqlNDe3VXxNFKJB/Xz9PMcDFPSMGsqNhIjsTd/sYUCA6bG34dS8s5C9U9EdATzhTASdwmMX6dAE6iXec4T8pBBg/7m3H7urMcjd+Cip9L1kwMrDJAlPi3vgsq1DhEZcE9+hM1GBJbmAV7zDLZrZULJ/1X1D8Y47LlK9NxyZIlp0Sj0SELFy78lzeeGmVhYaEEWe8tda4o0CMUoDZIBY0ePdoh8LzllluW27Y9E4N8bl5e3upmkhIBCu2DhHmNEUvXp5c5oT8wjvT+xHSY81BW1J6pD3cdblCuXRsINGmNuoFtcBiWpBYlcDu4e252bCscD051ttiuJQIRn7zRE6LjfPDSZy+3uT0XCtrjnChWg0XD9GKCRSPOrrOV76LnaUQlg39NrJ4iFadpvozhKm/f/AkbRRkx
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"
alt="Fig 1. Cassandra Streaming" style="float:
right;margin-right: 7px;margin-top: 7px;" /> To get an understanding of
the impact of these changes, letâs first have a look at the current streaming
code path. The diagram
below illustrates the stream session setup when a node attempts to stream data
from a peer. Letâs say, we have a 3 node cluster (Nodes A, B, C). Node C is
being rebuilt and has to stream all data that it is responsible for from A
&amp; B. C setups a streaming session with each of itâs peers (See: <a
href="https://issues.apache.org/jira/browse/CASSANDRA-4650">CASSANDRA-4560</a>
how Cassandra applies <a
href="https://en.wikipedia.org/wiki/Ford%E2%80%93Fulkerson_algorithm">Ford
Fulkerson</a> to optimize streaming peers). It exchanges messages to
request ranges and begins streaming data from the selected nodes.</p>
+
+<p>During the streaming phase, A collects all SSTables that have
partitions in the requested ranges. It streams each SSTable by serializing
individual partitions. Upon receiving the partition, node C reifies the data in
memory and then writes it to disk. This is necessary to accurately transfer
partitions from all possible SSTables for the requested ranges. This streaming
path generates garbage and could be avoided in scenarios where all partitions
within the SSTable need to be transmitted. This is common when youâre using
LeveledCompactionStrategy or have enabled partitioning SSTables by token range
(See: <a
href="http://issues.apache.org/jira/browse/CASSANDRA-6696">CASSANDRA-6696</a>),
etc.</p>
+
+<p>To solve this problem <a
href="http://issues.apache.org/jira/browse/CASSANDRA-14556">CASSANDRA-14556</a>
adds a Zero Copy streaming path. This significantly speeds up the transfer of
SSTables and reduces garbage and unnecessary object creation. It modifies the
streaming path to add additional information into the streaming header and uses
ZeroCopy APIs to transfer bytes to and from the network and disk. So now, an
SSTable may be transferred using this strategy when Cassandra detects that a
complete SSTable needs to be transferred.</p>
+
+<h2 id="how-do-i-use-this-feature">How do I use this
feature?</h2>
+
+<p>It just works. This feature is controlled using <code
class="highlighter-rouge">stream_entire_sstables</code> in
<code class="highlighter-rouge">cassandra.yaml</code> and
is enabled by default. Even though this feature is enabled, it will respect the
throttling limits as defined by <code
class="highlighter-rouge">stream_throughput_outbound_megabits_per_sec</code>.</p>
+
+<h2 id="impact">Impact</h2>
+
+<p>Cassandra can stream SSTables only bounded by the hardware
limitations (Network and Disk IO). With this optimization, we hope to make
Cassandra more performant and reliable.</p>
+
+<p>Microbenchmarking this feature shows a marked improvement (higher is
better). Block Stream Writers are the ZeroCopy writers and Partial Stream
Writers are the existing writers.</p>
+
+<table class="table-condensed table-bordered table-hover">
+ <thead>
+ <tr>
+ <th>Benchmark</th>
+ <th>Mode</th>
+ <th>Cnt</th>
+ <th>Score</th>
+ <th>Error</th>
+ <th>Units</th>
+ </tr>
+ </thead>
+ <tbody>
+ <tr>
+ <td>ZeroCopyStreamingBenchmark.blockStreamReader</td>
+ <td>thrpt</td>
+ <td>10</td>
+ <td>20.119</td>
+ <td>± 1.300</td>
+ <td>ops/s</td>
+ </tr>
+ <tr>
+ <td>ZeroCopyStreamingBenchmark.blockStreamWriter</td>
+ <td>thrpt</td>
+ <td>10</td>
+ <td>1339.672</td>
+ <td>± 352.242</td>
+ <td>ops/s</td>
+ </tr>
+ <tr>
+ <td>ZeroCopyStreamingBenchmark.partialStreamReader</td>
+ <td>thrpt</td>
+ <td>10</td>
+ <td>0.590</td>
+ <td>± 0.135</td>
+ <td>ops/s</td>
+ </tr>
+ <tr>
+ <td>ZeroCopyStreamingBenchmark.partialStreamWriter</td>
+ <td>thrpt</td>
+ <td>10</td>
+ <td>17.556</td>
+ <td>± 0.323</td>
+ <td>ops/s</td>
+ </tr>
+ </tbody>
+</table>
+
+<h2 id="conclusion">Conclusion</h2>
+
+<p>If youâre a Cassandra user, we would love to hear back from you.
Please send us feedback via user <a
href="http://cassandra.apache.org/community/">Mailing
List</a>, <a
href="https://issues.apache.org/jira/projects/CASSANDRA/summary">Jira</a>,
or <a
href="http://cassandra.apache.org/community/">IRC</a> (or
any combination of the three).</p></content><author><name>The Apache
Cassandra Community</name></author><summary type="html">Streaming in Apache
Cassandra powers host replacement, range movements, and cluster expansions.
Streaming plays a crucial role in the cluster and as such its performance is
key to not only the speed of the operations its used in but the clusterâs
health generally. In Apache Cassandra 4.0, we have introduced an improved
streaming implementation that reduces GC pressure and increases throughput
several folds and are now limited, in some cases, only by the disk / network I
O (See: CASSANDRA-14556).</summary></entry></feed>
\ No newline at end of file
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]