Squashed commit of the following:
commit f55a17f71ec97513a6968b1ea3c359bc6238cc5e
Author: Yakov Zhdanov <yzhda...@gridgain.com>
Date: Fri Jul 31 13:32:32 2015 +0300
review
commit 58ca345f622dbadfba7ef2d3dce850c4baa1f319
Merge: 5f921f6 7ed4d15
Author: Yakov Zhdanov <yzhda...@gridgain.com>
Date: Fri Jul 31 13:24:51 2015 +0300
Merge branches 'ignite-752-2' and 'master' of
https://git-wip-us.apache.org/repos/asf/incubator-ignite into ignite-752-2
commit 5f921f62dd6563a88b2ecdde92a2b2ee8218ec95
Author: Denis Magda <dma...@gridgain.com>
Date: Wed Jul 29 10:40:44 2015 +0300
ignite-752-2: added info on the lowest failure detection timeout to the
documentation
commit 55f0eb56967d2cc9bdf62c3fb665521a59ddaf33
Author: Denis Magda <dma...@gridgain.com>
Date: Wed Jul 29 09:15:29 2015 +0300
ignite-752-2: supported connection check frequency even for cases when
failure timeout is ignored; performance optimizations
Project: http://git-wip-us.apache.org/repos/asf/incubator-ignite/repo
Commit: http://git-wip-us.apache.org/repos/asf/incubator-ignite/commit/44072f80
Tree: http://git-wip-us.apache.org/repos/asf/incubator-ignite/tree/44072f80
Diff: http://git-wip-us.apache.org/repos/asf/incubator-ignite/diff/44072f80
Branch: refs/heads/ignite-946
Commit: 44072f806d8d14d716475a3665d0afdf004c6db2
Parents: 7ed4d15
Author: Denis Magda <dma...@gridgain.com>
Authored: Fri Jul 31 13:35:46 2015 +0300
Committer: Yakov Zhdanov <yzhda...@gridgain.com>
Committed: Fri Jul 31 13:35:46 2015 +0300
----------------------------------------------------------------------
.../ignite/spi/discovery/tcp/ServerImpl.java | 42 +++++++++++---------
.../spi/discovery/tcp/TcpDiscoverySpi.java | 2 +-
2 files changed, 24 insertions(+), 20 deletions(-)
----------------------------------------------------------------------
http://git-wip-us.apache.org/repos/asf/incubator-ignite/blob/44072f80/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/ServerImpl.java
----------------------------------------------------------------------
diff --git
a/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/ServerImpl.java
b/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/ServerImpl.java
index 547347c..47ba8e6 100644
---
a/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/ServerImpl.java
+++
b/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/ServerImpl.java
@@ -1787,6 +1787,9 @@ class ServerImpl extends TcpDiscoveryImpl {
/** Connection check frequency. */
private long connCheckFreq;
+ /** Connection check threshold. */
+ private long connCheckThreshold;
+
/**
*/
protected RingMessageWorker() {
@@ -1799,19 +1802,22 @@ class ServerImpl extends TcpDiscoveryImpl {
* Initializes connection check frequency. Used only when failure
detection timeout is enabled.
*/
private void initConnectionCheckFrequency() {
- if (spi.failureDetectionTimeoutEnabled()) {
- for (int i = 3; i > 0; i--) {
- connCheckFreq = spi.failureDetectionTimeout() / i;
-
- if (connCheckFreq > 0)
- break;
- }
+ if (spi.failureDetectionTimeoutEnabled())
+ connCheckThreshold = spi.failureDetectionTimeout();
+ else
+ connCheckThreshold = Math.min(spi.getSocketTimeout(),
spi.getHeartbeatFrequency());
- assert connCheckFreq > 0;
+ for (int i = 3; i > 0; i--) {
+ connCheckFreq = connCheckThreshold / i;
- if (log.isDebugEnabled())
- log.debug("Connection check frequency is calculated: " +
connCheckFreq);
+ if (connCheckFreq > 10)
+ break;
}
+
+ assert connCheckFreq > 0;
+
+ if (log.isDebugEnabled())
+ log.debug("Connection check frequency is calculated: " +
connCheckFreq);
}
/**
@@ -2306,9 +2312,9 @@ class ServerImpl extends TcpDiscoveryImpl {
// If node existed on connection initialization we
should check
// whether it has not gone yet.
- if (nextNodeExists && pingNode(next))
- U.error(log, "Failed to send message to next
node [msg=" + msg +
- ", next=" + next + ']', err);
+ if (nextNodeExists)
+ U.warn(log, "Failed to send message to next
node [msg=" + msg + ", next=" + next +
+ ", errMsg=" + (err != null ?
err.getMessage() : "N/A") + ']');
else if (log.isDebugEnabled())
log.debug("Failed to send message to next node
[msg=" + msg + ", next=" + next +
", errMsg=" + (err != null ?
err.getMessage() : "N/A") + ']');
@@ -4025,7 +4031,7 @@ class ServerImpl extends TcpDiscoveryImpl {
/**
* Check the last time a heartbeat message received. If the time is
bigger than {@code hbCheckTimeout} than
- * {@link TcpDiscoveryStatusCheckMessage} is sent accros the ring.
+ * {@link TcpDiscoveryStatusCheckMessage} is sent across the ring.
*/
private void checkHeartbeatsReceiving() {
if (lastTimeStatusMsgSent < locNode.lastUpdateTime())
@@ -4045,11 +4051,9 @@ class ServerImpl extends TcpDiscoveryImpl {
* Check connection aliveness status.
*/
private void checkConnection() {
- if (!spi.failureDetectionTimeoutEnabled())
- return;
-
- if (!failureThresholdReached && U.currentTimeMillis() -
locNode.lastDataReceivedTime()
- >= spi.failureDetectionTimeout() && ring.hasRemoteNodes() &&
spiStateCopy() == CONNECTED) {
+ if (spi.failureDetectionTimeoutEnabled() &&
!failureThresholdReached &&
+ U.currentTimeMillis() - locNode.lastDataReceivedTime() >=
connCheckThreshold &&
+ ring.hasRemoteNodes() && spiStateCopy() == CONNECTED) {
log.info("Local node seems to be disconnected from topology
(failure detection timeout " +
"is reached): [failureDetectionTimeout=" +
spi.failureDetectionTimeout() +
http://git-wip-us.apache.org/repos/asf/incubator-ignite/blob/44072f80/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/TcpDiscoverySpi.java
----------------------------------------------------------------------
diff --git
a/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/TcpDiscoverySpi.java
b/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/TcpDiscoverySpi.java
index 09690dc..3216166 100644
---
a/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/TcpDiscoverySpi.java
+++
b/modules/core/src/main/java/org/apache/ignite/spi/discovery/tcp/TcpDiscoverySpi.java
@@ -74,7 +74,7 @@ import java.util.concurrent.atomic.*;
* {@link IgniteConfiguration#setFailureDetectionTimeout(long)}. This failure
timeout automatically controls the
* following parameters: {@link #getSocketTimeout()}, {@link
#getAckTimeout()}, {@link #getMaxAckTimeout()},
* {@link #getReconnectCount()}. If any of those parameters is set explicitly,
then the failure timeout setting will be
- * ignored.
+ * ignored. As an example, for stable low-latency networks the failure
detection timeout may be set to ~120 ms.
* <p>
* If it's required to perform advanced settings of failure detection and
* {@link IgniteConfiguration#getFailureDetectionTimeout()} is unsuitable then
various {@code TcpDiscoverySpi}
