[
https://issues.apache.org/jira/browse/CLOUDSTACK-9107?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15041413#comment-15041413
]
ASF GitHub Bot commented on CLOUDSTACK-9107:
--------------------------------------------
Github user DaanHoogland commented on a diff in the pull request:
https://github.com/apache/cloudstack/pull/1172#discussion_r46669997
--- Diff:
engine/orchestration/src/com/cloud/agent/manager/ClusteredAgentManagerImpl.java
---
@@ -1411,14 +1411,12 @@ protected void runInContext() {
final double managedHostsCount =
allManagedRoutingAgents.size();
if (allHostsCount > 0.0) {
final double load = managedHostsCount /
allHostsCount;
- if (load >= ConnectedAgentThreshold.value()) {
- s_logger.debug("Scheduling agent
rebalancing task as the average agent load " + load + " is more than the
threshold " +
- ConnectedAgentThreshold.value());
+ if (load > ConnectedAgentThreshold.value()) {
+ s_logger.debug("Scheduling agent
rebalancing task as the average agent load " + load + " is more than the
threshold " + ConnectedAgentThreshold.value());
scheduleRebalanceAgents();
_agentLbHappened = true;
} else {
- s_logger.debug("Not scheduling agent
rebalancing task as the averages load " + load + " is less than the threshold "
+
- ConnectedAgentThreshold.value());
+ s_logger.debug("Not scheduling agent
rebalancing task as the average load " + load + " has not crossed the threshold
" + ConnectedAgentThreshold.value());
--- End diff --
I can see teh '>=' -> '>' change in the line above. Has this line any
logical changes as well?
> Description of agent.load.threshold and the actual behavior are not in sync
> ---------------------------------------------------------------------------
>
> Key: CLOUDSTACK-9107
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9107
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Components: Management Server
> Affects Versions: 4.5.0, 4.6.0
> Reporter: Koushik Das
> Assignee: Koushik Das
> Fix For: 4.7.0
>
>
> Global setting to control agent load balancing threshold
> "agent.load.threshold" has a description as "What percentage of the agents
> can be held by one management server before load balancing happens" .
> When the above setting is set to a value of 1, the behavior in logs is
> "Scheduling agent rebalancing task as the average agent load 1.0 is more than
> the threshold 1.0"
> The ambiguity if re-balancing should happen or not can be fixed either in the
> description or in the actual behavior.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
