aboutsummaryrefslogtreecommitdiffstats
path: root/lib
diff options
context:
space:
mode:
authorSverker Eriksson <[email protected]>2018-07-18 16:52:41 +0200
committerSverker Eriksson <[email protected]>2018-07-18 16:52:41 +0200
commit4f2fa2edabcc5807a879a6f60ecb2f28d5ad6085 (patch)
treedc1a2141ac4a9c79563f1ac5890e9f75d798dec4 /lib
parent6a98a4b05ae77739d27f2a9fb018786906e6918c (diff)
downloadotp-4f2fa2edabcc5807a879a6f60ecb2f28d5ad6085.tar.gz
otp-4f2fa2edabcc5807a879a6f60ecb2f28d5ad6085.tar.bz2
otp-4f2fa2edabcc5807a879a6f60ecb2f28d5ad6085.zip
kernel: Improve net_ticktime docs
Diffstat (limited to 'lib')
-rw-r--r--lib/kernel/doc/src/kernel_app.xml37
1 files changed, 22 insertions, 15 deletions
diff --git a/lib/kernel/doc/src/kernel_app.xml b/lib/kernel/doc/src/kernel_app.xml
index 3914226a3e..ccb7d17d25 100644
--- a/lib/kernel/doc/src/kernel_app.xml
+++ b/lib/kernel/doc/src/kernel_app.xml
@@ -310,24 +310,31 @@
<tag><c>net_ticktime = TickTime</c></tag>
<item>
<marker id="net_ticktime"></marker>
- <p>Specifies the <c>net_kernel</c> tick time. <c>TickTime</c>
- is specified in seconds. Once every <c>TickTime/4</c> second, all
- connected nodes are ticked (if anything else is written
- to a node). If nothing is received from another node
- within the last four tick times, that node is considered
- to be down. This ensures that nodes that are not responding,
- for reasons such as hardware errors, are considered to be
- down.</p>
- <p>The time <c>T</c>, in which a node that is not responding is
- detected, is calculated as <c><![CDATA[MinT < T < MaxT]]></c>, where:</p>
+ <p>Specifies the <c>net_kernel</c> tick time in seconds. This is the
+ approximate time a connected node may be unresponsive until it is
+ considered down and thereby disconnected.</p>
+ <p>Once every <c>TickTime/4</c> seconds, each connected node is ticked
+ if nothing has been sent to it during that last <c>TickTime/4</c>
+ interval. A tick is a small package sent on the connection. A connected
+ node is considered to be down if no ticks or payload packages have been
+ received during the last four <c>TickTime/4</c> intervals. This ensures
+ that nodes that are not responding, for reasons such as hardware errors,
+ are considered to be down.</p>
+ <p>As the availability is only checked every <c>TickTime/4</c> seconds,
+ the actual time <c>T</c> a node have been unresponsive when
+ detected may vary between <c>MinT</c> and <c>MaxT</c>,
+ where:</p>
<code type="none">
MinT = TickTime - TickTime / 4
MaxT = TickTime + TickTime / 4</code>
- <p><c>TickTime</c> defaults to <c>60</c> (seconds). Thus,
- <c><![CDATA[45 < T < 75]]></c> seconds.</p>
- <p>Notice that <em>all</em> communicating nodes are to have the <em>same</em>
- <c>TickTime</c> value specified.</p>
- <p>Normally, a terminating node is detected immediately.</p>
+ <p><c>TickTime</c> defaults to <c>60</c> seconds. Thus,
+ <c><![CDATA[45 < T < 75]]></c> seconds.</p>
+ <p>Notice that <em>all</em> communicating nodes are to have the
+ <em>same</em> <c>TickTime</c> value specified, as it determines both the
+ frequency of outgoing ticks and the expected frequency of incominging
+ ticks.</p>
+ <p>Normally, a terminating node is detected immediately by the transport
+ protocol (like TCP/IP).</p>
</item>
<tag><c>shutdown_timeout = integer() | infinity</c></tag>
<item>