[JBoss JIRA] (MODCLUSTER-607) Double decay factor
by Radoslav Husar (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-607?page=com.atlassian.jira.pl... ]
Work on MODCLUSTER-607 started by Radoslav Husar.
-------------------------------------------------
> Double decay factor
> -------------------
>
> Key: MODCLUSTER-607
> URL: https://issues.jboss.org/browse/MODCLUSTER-607
> Project: mod_cluster
> Issue Type: Enhancement
> Components: Core & Container Integration (Java)
> Affects Versions: 1.2.13.Final, 1.3.7.Final
> Reporter: Leandro Quiroga
> Assignee: Radoslav Husar
> Priority: Minor
> Labels: decay
>
> The int type is too restrictive for specifying the decay factor.
> Using a decay factor of 1 means no decay.
> Next value is 2, and the importance of loads decreases too fast:
> ||Instant||Importance||
> |0|100%|
> |1|50%|
> |2|25%|
> |3|12.5%|
> |4|6.3%|
> |5|3.1%|
> |6|1.6%|
> |7|0.8%|
> |8|0.4%|
> |9|0.2%|
> Next value is 3, and the importance of loads decreases even faster:
> ||Instant||Importance||
> |0|100%|
> |1|33%|
> |2|11%|
> |3|3.7%|
> |4|1.2%|
> |5|0.4%|
> |6|0.1%|
> |7|0.0%|
> |8|0.0%|
> |9|0.0%|
>
> Using a double for decay factor isn't complex and allows much more precision.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-607) Double decay factor
by Radoslav Husar (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-607?page=com.atlassian.jira.pl... ]
Radoslav Husar reassigned MODCLUSTER-607:
-----------------------------------------
Assignee: Radoslav Husar (was: Jean-Frederic Clere)
> Double decay factor
> -------------------
>
> Key: MODCLUSTER-607
> URL: https://issues.jboss.org/browse/MODCLUSTER-607
> Project: mod_cluster
> Issue Type: Enhancement
> Components: Core & Container Integration (Java)
> Affects Versions: 1.2.13.Final, 1.3.7.Final
> Reporter: Leandro Quiroga
> Assignee: Radoslav Husar
> Priority: Minor
> Labels: decay
>
> The int type is too restrictive for specifying the decay factor.
> Using a decay factor of 1 means no decay.
> Next value is 2, and the importance of loads decreases too fast:
> ||Instant||Importance||
> |0|100%|
> |1|50%|
> |2|25%|
> |3|12.5%|
> |4|6.3%|
> |5|3.1%|
> |6|1.6%|
> |7|0.8%|
> |8|0.4%|
> |9|0.2%|
> Next value is 3, and the importance of loads decreases even faster:
> ||Instant||Importance||
> |0|100%|
> |1|33%|
> |2|11%|
> |3|3.7%|
> |4|1.2%|
> |5|0.4%|
> |6|0.1%|
> |7|0.0%|
> |8|0.0%|
> |9|0.0%|
>
> Using a double for decay factor isn't complex and allows much more precision.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-587) Docs: Undertow hosts root location is exposed(by default) to mod_cluster load balancer
by Bogdan Sikora (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-587?page=com.atlassian.jira.pl... ]
Bogdan Sikora updated MODCLUSTER-587:
-------------------------------------
Tester: Jan Kašík (was: Bogdan Sikora)
> Docs: Undertow hosts root location is exposed(by default) to mod_cluster load balancer
> --------------------------------------------------------------------------------------
>
> Key: MODCLUSTER-587
> URL: https://issues.jboss.org/browse/MODCLUSTER-587
> Project: mod_cluster
> Issue Type: Bug
> Components: Documentation & Demos
> Reporter: Bogdan Sikora
> Assignee: Michal Karm Babacek
>
> Undertow can be configured to serve static content per host via the location resource. This location resource is exposed to the mod_cluster balancer.
> {noformat}
> "context" => {
> "/custom_location" => {
> "requests" => 0,
> "status" => "enabled"
> },
> {noformat}
> However, all EAP standalone profiles have enabled root location by default. This root location is then exposed to the balancer( Application root "/" is registered).
> {noformat}
> "context" => {
> "/" => {
> "requests" => 0,
> "status" => "enabled"
> },
> {noformat}
> Root application matches with any application call and therefore mod_cluster balancer is unable to correctly route requests.
> To make mod_cluster work, one must delete root location from worker nodes.
> {noformat}
> /subsystem=undertow/server=default-server/host=default-host/location=\/:remove()
> {noformat}
> or excluding context ROOT
> {noformat}
> /subsystem=modcluster/mod-cluster-config=configuration:write-attribute(name=excluded-contexts,value="ROOT")
> {noformat}
> *_+Proposal:+_*
> Exclude ROOT context by default
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-596) mod_cluster stop/stop-context operations do not send STOP-* in when session draining was unsuccessful
by Bogdan Sikora (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-596?page=com.atlassian.jira.pl... ]
Bogdan Sikora updated MODCLUSTER-596:
-------------------------------------
Tester: (was: Bogdan Sikora)
> mod_cluster stop/stop-context operations do not send STOP-* in when session draining was unsuccessful
> -----------------------------------------------------------------------------------------------------
>
> Key: MODCLUSTER-596
> URL: https://issues.jboss.org/browse/MODCLUSTER-596
> Project: mod_cluster
> Issue Type: Bug
> Components: Core & Container Integration (Java)
> Affects Versions: 1.3.6.Final
> Reporter: Radoslav Husar
> Assignee: Radoslav Husar
> Fix For: 2.0.0.Alpha1, 1.3.7.Final
>
>
> CLI operation description
> {noformat}
> [standalone@localhost:9990 /] /subsystem=modcluster/:read-operation-description(name=stop
> {
> "outcome" => "success",
> "result" => {
> "operation-name" => "stop",
> "description" => "Tell reverse proxies that all contexts on the node can't process requests.",
> "request-properties" => {"waittime" => {
> "type" => INT,
> "description" => "Timeout to wait for all contexts to stop.",
> "expressions-allowed" => false,
> "required" => false,
> "nillable" => true,
> "default" => 10,
> "unit" => "SECONDS"
> }},
> "reply-properties" => {},
> "read-only" => false,
> "runtime-only" => true
> }
> }
> {noformat}
> With default and non-distributable context or draining strategy set to ALWAYS.
> * stop() or stop-context is called
> * Disable creating any new session by sending DISABLE-APP command to balancer
> {noformat}
> "context" => {"/clusterbench" => {
> "requests" => 0,
> "status" => "disabled"
> }}
> {noformat}
> * Session drain
> {noformat}
> 2017-05-10 09:19:09,956 INFO [org.jboss.modcluster] (management-handler-thread - 1) MODCLUSTER000046: Starting to drain 1 active sessions from default-host:/clusterbench in 10 seconds.
> {noformat}
> * Session timeout is 5 minutes so it fails
> {noformat}
> 2017-05-10 09:26:05,081 WARN [org.jboss.modcluster] (management-handler-thread - 1) MODCLUSTER000025: Failed to drain 1 remaining active sessions from default-host:/clusterbench within 10.0 seconds
> {noformat}
> *_+{color:red}Issue{color}+_*
> * Stop-app should be sent to balancer to stop context/node, but node/context stays disabled. This happens even if session drain is successful.
> {noformat}
> "context" => {"/clusterbench" => {
> "requests" => 0,
> "status" => "disabled"
> }}
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-608) failonstatus parameter ignored
by Bogdan Sikora (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-608?page=com.atlassian.jira.pl... ]
Bogdan Sikora moved JBCS-381 to MODCLUSTER-608:
-----------------------------------------------
Project: mod_cluster (was: JBoss Core Services)
Key: MODCLUSTER-608 (was: JBCS-381)
Workflow: GIT Pull Request workflow (was: CDW v1)
Component/s: Native (httpd modules)
(was: mod_cluster)
Affects Version/s: 1.3.5.Final
(was: httpd 2.4.23 GA)
(was: httpd 2.4.23 SP1 GA)
> failonstatus parameter ignored
> ------------------------------
>
> Key: MODCLUSTER-608
> URL: https://issues.jboss.org/browse/MODCLUSTER-608
> Project: mod_cluster
> Issue Type: Bug
> Components: Native (httpd modules)
> Affects Versions: 1.3.5.Final
> Reporter: Bogdan Sikora
> Assignee: Michal Karm Babacek
>
> Setting failonstatus to a specific code should result in the error state of the worker that returned the code.
> *Example:*
> Httpd mod_cluster balancer with 2 tomcat workers that have deployed clusterbench app
> {noformat}
> ProxyPass / balancer://balancerxxx stickysession=JSESSIONID|jsessionid nofailover=on failonstatus=201
> ProxyPassReverse / balancer://balancerxxx
> ProxyPreserveHost on
> {noformat}
> {noformat}
> Checking for http://192.168.122.206:2080/clusterbench/jvmroute response code: was 200, expected [200, 201, 202, 203]
> {noformat}
> Let's send special request that returns 201(failonstatus code)
> {noformat}
> Checking for http://192.168.122.206:2080/clusterbench/httpcode/?http_code=201 response code: was 201, expected 201
> {noformat}
> This should result in worker in error state
> {noformat}
> <h1> Node tomcat-7-1 (ajp://192.168.122.206:8009): </h1>
> ...Status: NOTOK,Elected: 13,Read: 245,Transferred: 0,Connected: 0,Load: 100...
> ><h1> Node tomcat-7-2 (ajp://192.168.122.206:8110): </h1>
> ...Status: OK,Elected: 1,Read: 10,Transferred: 0,Connected: 0,Load: 100...
> {noformat}
> as in example. This example is made with 2.4.6 httpd.
> Debug log
> {noformat}
> [Wed Jul 26 09:13:12.646541 2017] [:error] [pid 22223] [client 192.168.122.206:41482] balancer://balancerxxx: Forcing worker (ajp://192.168.122.206:8009) into error state due to status code 201 matching 'failonstatus' balancer parameter
> {noformat}
> *Issue:*
> 2.4.23+ Httpd ignores failonstatus parameter.
> Return of 201 as in example trigger no error state.
> There is no mention in the log and workers on the console are in OK state.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-606) Documentation Error - Chapter 10 - Time decay formulas
by Leandro Quiroga (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-606?page=com.atlassian.jira.pl... ]
Leandro Quiroga updated MODCLUSTER-606:
---------------------------------------
Description:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) */* (∑^H^~i=0~ *1/* D ^i^)
... where D = decayFactor, and H = history.
was:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * / * (∑^H^~i=0~ *1/* D ^i^)
... where D = decayFactor, and H = history.
> Documentation Error - Chapter 10 - Time decay formulas
> ------------------------------------------------------
>
> Key: MODCLUSTER-606
> URL: https://issues.jboss.org/browse/MODCLUSTER-606
> Project: mod_cluster
> Issue Type: Bug
> Components: Documentation & Demos
> Affects Versions: 1.2.13.Final, 1.3.7.Final
> Reporter: Leandro Quiroga
> Assignee: Jean-Frederic Clere
> Priority: Minor
> Labels: documentation
> Original Estimate: 1 hour
> Remaining Estimate: 1 hour
>
> In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
> ... where D = decayFactor, and H = history.
>
> But both formulas are wrong.
> It should says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) */* (∑^H^~i=0~ *1/* D ^i^)
> ... where D = decayFactor, and H = history.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-606) Documentation Error - Chapter 10 - Time decay formulas
by Leandro Quiroga (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-606?page=com.atlassian.jira.pl... ]
Leandro Quiroga updated MODCLUSTER-606:
---------------------------------------
Description:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * / * (∑^H^~i=0~ *1/* D ^i^)
... where D = decayFactor, and H = history.
was:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) */* (∑^H^~i=0~ *1/*D ^i^)
... where D = decayFactor, and H = history.
> Documentation Error - Chapter 10 - Time decay formulas
> ------------------------------------------------------
>
> Key: MODCLUSTER-606
> URL: https://issues.jboss.org/browse/MODCLUSTER-606
> Project: mod_cluster
> Issue Type: Bug
> Components: Documentation & Demos
> Affects Versions: 1.2.13.Final, 1.3.7.Final
> Reporter: Leandro Quiroga
> Assignee: Jean-Frederic Clere
> Priority: Minor
> Labels: documentation
> Original Estimate: 1 hour
> Remaining Estimate: 1 hour
>
> In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
> ... where D = decayFactor, and H = history.
>
> But both formulas are wrong.
> It should says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) * / * (∑^H^~i=0~ *1/* D ^i^)
> ... where D = decayFactor, and H = history.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-606) Documentation Error - Chapter 10 - Time decay formulas
by Leandro Quiroga (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-606?page=com.atlassian.jira.pl... ]
Leandro Quiroga updated MODCLUSTER-606:
---------------------------------------
Description:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) */* (∑^H^~i=0~ *1/*D ^i^)
... where D = decayFactor, and H = history.
was:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~/D^0^ + L ~1~/D^1^ + L ~2~/D^2^ + L ~3~/D^3^ + ... + L ~H~/D ^H^) / (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) / (∑^H^~i=0~ 1/D ^i^)
... where D = decayFactor, and H = history.
> Documentation Error - Chapter 10 - Time decay formulas
> ------------------------------------------------------
>
> Key: MODCLUSTER-606
> URL: https://issues.jboss.org/browse/MODCLUSTER-606
> Project: mod_cluster
> Issue Type: Bug
> Components: Documentation & Demos
> Affects Versions: 1.2.13.Final, 1.3.7.Final
> Reporter: Leandro Quiroga
> Assignee: Jean-Frederic Clere
> Priority: Minor
> Labels: documentation
> Original Estimate: 1 hour
> Remaining Estimate: 1 hour
>
> In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
> ... where D = decayFactor, and H = history.
>
> But both formulas are wrong.
> It should says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~*/D ^0^* + L ~1~/D *^1^* + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) */ (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)*
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) */* (∑^H^~i=0~ *1/*D ^i^)
> ... where D = decayFactor, and H = history.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months
[JBoss JIRA] (MODCLUSTER-606) Documentation Error - Chapter 10 - Time decay formulas
by Leandro Quiroga (JIRA)
[ https://issues.jboss.org/browse/MODCLUSTER-606?page=com.atlassian.jira.pl... ]
Leandro Quiroga updated MODCLUSTER-606:
---------------------------------------
Description:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~/D^0^ + L ~1~/D^1^ + L ~2~/D^2^ + L ~3~/D^3^ + ... + L ~H~/D ^H^) / (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L ~i~/D ^i^) / (∑^H^~i=0~ 1/D ^i^)
... where D = decayFactor, and H = history.
was:
In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
... or more concisely as:
L = (∑^H^~i=0~ L~i~/D^i^) * (∑^H^~i=0~ D^i^)
... where D = decayFactor, and H = history.
But both formulas are wrong.
It should says:
The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
L = (L0/D0 + L1/D1 + L2/D2 + L3/D3 + ... + LH/DH) / (1/D0 + 1/D1 + 1/D2 + 1/D3 + ... 1/DH)
... or more concisely as:
L = (∑Hi=0 Li/Di) / (∑Hi=0 1/Di)
... where D = decayFactor, and H = history.
> Documentation Error - Chapter 10 - Time decay formulas
> ------------------------------------------------------
>
> Key: MODCLUSTER-606
> URL: https://issues.jboss.org/browse/MODCLUSTER-606
> Project: mod_cluster
> Issue Type: Bug
> Components: Documentation & Demos
> Affects Versions: 1.2.13.Final, 1.3.7.Final
> Reporter: Leandro Quiroga
> Assignee: Jean-Frederic Clere
> Priority: Minor
> Labels: documentation
> Original Estimate: 1 hour
> Remaining Estimate: 1 hour
>
> In chapter 10 (http://docs.jboss.org/mod_cluster/1.3.0/html_single/#java.load), it says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~ + L ~1~/D + L ~2~/D ^2^ + L ~3~/D ^3^ + ... + L ~H~/D ^H^) * (1 + D + D ^2^ + D ^3^ + ... D ^H^)
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) * (∑^H^~i=0~ D ^i^)
> ... where D = decayFactor, and H = history.
>
> But both formulas are wrong.
> It should says:
> The DynamicLoadBalanceFactorProvider applies a time decay function to the loads returned by each metric. The aggregate load, with respect to previous load values, can be expressed by the following formula:
> L = (L ~0~/D^0^ + L ~1~/D^1^ + L ~2~/D^2^ + L ~3~/D^3^ + ... + L ~H~/D ^H^) / (1/D ^0^ + 1/D ^1^ + 1/D ^2^ + 1/D ^3^ + ... 1/D ^H^)
> ... or more concisely as:
> L = (∑^H^~i=0~ L ~i~/D ^i^) / (∑^H^~i=0~ 1/D ^i^)
> ... where D = decayFactor, and H = history.
--
This message was sent by Atlassian JIRA
(v7.2.3#72005)
7 years, 4 months