/usr/share/doc/drbd-doc/users-guide/ch-latency.html is in drbd-doc 8.4~20151102-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | <?xml version="1.0" encoding="UTF-8" standalone="no"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>Chapter 16. Optimizing DRBD latency</title><link rel="stylesheet" type="text/css" href="default.css" /><meta name="generator" content="DocBook XSL Stylesheets V1.79.1" /><link rel="home" href="drbd-users-guide.html" title="The DRBD User’s Guide" /><link rel="up" href="p-performance.html" title="Part V. Optimizing DRBD performance" /><link rel="prev" href="s-throughput-tuning.html" title="15.3. Tuning recommendations" /><link rel="next" href="s-latency-overhead-expectations.html" title="16.2. Latency overhead expectations" /></head><body><div class="navheader"><table width="100%" summary="Navigation header"><tr><th colspan="3" align="center">Chapter 16. Optimizing DRBD latency</th></tr><tr><td width="20%" align="left"><a accesskey="p" href="s-throughput-tuning.html">Prev</a> </td><th width="60%" align="center">Part V. Optimizing DRBD performance</th><td width="20%" align="right"> <a accesskey="n" href="s-latency-overhead-expectations.html">Next</a></td></tr></table><hr /></div><div class="chapter"><div class="titlepage"><div><div><h2 class="title"><a id="ch-latency"></a>Chapter 16. Optimizing DRBD latency</h2></div></div></div><div class="toc"><p><strong>Table of Contents</strong></p><dl class="toc"><dt><span class="section"><a href="ch-latency.html#s-latency-hardware">16.1. Hardware considerations</a></span></dt><dt><span class="section"><a href="s-latency-overhead-expectations.html">16.2. Latency overhead expectations</a></span></dt><dt><span class="section"><a href="s-latency-tuning.html">16.3. Tuning recommendations</a></span></dt><dd><dl><dt><span class="section"><a href="s-latency-tuning.html#s-latency-tuning-cpu-mask">16.3.1. Setting DRBD’s CPU mask</a></span></dt><dt><span class="section"><a href="s-latency-tuning.html#s-latency-tuning-mtu-size">16.3.2. Modifying the network MTU</a></span></dt><dt><span class="section"><a href="s-latency-tuning.html#s-latency-tuning-deadline-scheduler">16.3.3. Enabling the <code class="literal">deadline</code> I/O scheduler</a></span></dt></dl></dd></dl></div><p>This chapter deals with optimizing DRBD latency. It examines some
hardware considerations with regard to latency minimization, and
details tuning recommendations for that purpose.</p><div class="section"><div class="titlepage"><div><div><h2 class="title" style="clear: both"><a id="s-latency-hardware"></a>16.1. Hardware considerations</h2></div></div></div><p>DRBD latency is affected by both the latency of the underlying I/O
subsystem (disks, controllers, and corresponding caches), and the
latency of the replication network.</p><p><strong>I/O subsystem latency. </strong><a id="idp46951088" class="indexterm"></a>I/O subsystem latency is primarily a function of
disk rotation speed. Thus, using fast-spinning disks is a valid
approach for reducing I/O subsystem latency.</p><p>Likewise, the use of a <a id="idp46952816" class="indexterm"></a>
battery-backed write cache (BBWC) reduces write completion times, also
reducing write latency. Most reasonable storage subsystems come with
some form of battery-backed cache, and allow the administrator to
configure which portion of this cache is used for read and write
operations. The recommended approach is to disable the disk read cache
completely and use all cache memory available for the disk write
cache.</p><p><strong>Network latency. </strong><a id="idp46955296" class="indexterm"></a>Network latency is, in essence, the packet
round-trip time ( ) between hosts. It is influenced by a number of
factors, most of which are irrelevant on the dedicated, back-to-back
network connections recommended for use as DRBD replication
links. Thus, it is sufficient to accept that a certain amount of
latency always exists in Gigabit Ethernet links, which typically is on
the order of 100 to 200 microseconds (μs) packet RTT.</p><p>Network latency may typically be pushed below this limit only by using
lower-latency network protocols, such as running DRBD over Dolphin
Express using Dolphin SuperSockets.</p></div></div><div class="navfooter"><hr /><table width="100%" summary="Navigation footer"><tr><td width="40%" align="left"><a accesskey="p" href="s-throughput-tuning.html">Prev</a> </td><td width="20%" align="center"><a accesskey="u" href="p-performance.html">Up</a></td><td width="40%" align="right"> <a accesskey="n" href="s-latency-overhead-expectations.html">Next</a></td></tr><tr><td width="40%" align="left" valign="top">15.3. Tuning recommendations </td><td width="20%" align="center"><a accesskey="h" href="drbd-users-guide.html">Home</a></td><td width="40%" align="right" valign="top"> 16.2. Latency overhead expectations</td></tr></table></div></body></html>
|