You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Dmitriy Govorukhin (JIRA)" <ji...@apache.org> on 2019/03/14 08:56:00 UTC
[jira] [Updated] (IGNITE-10945) Document Baseline auto-adjust
feature
[ https://issues.apache.org/jira/browse/IGNITE-10945?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Dmitriy Govorukhin updated IGNITE-10945:
----------------------------------------
Description:
From IGNITE-8571:
Now we have only one way to change BLAT - manually update it via console.sh or API.
We need to add the possibility to change it automatically. Adjust to current topology.
So, I propose 3 new parameters which would be responsible to tune this feature.
1. Flag autoAdjustEnabled - true/false. Easy. Manual baseline control or auto adjusting baseline.
2. autoAdjustTimeout - time which we would wait after the actual topology change. But it would be reset if new discovery event happened. (node join/exit).
3. autoAdjustMaxTimeout - time which we would wait from the first dicovery event in the chain. If we achieved it than we would change BLAT right away (no matter were another node join/exit happened or not).
We need to change API next way:
1. org.apache.ignite.IgniteCluster
*Add*
isBaselineAutoAdjustEnabled()
setBaselineAutoAdjustEnabled(boolean enabled);
setBaselineAutoAdjustTimeout(long timeoutInMs);
setBaselineAutoAdjustMaxTimeout(long timeoutInMs);
2. org.apache.ignite.configuration.IgniteConfiguration
*Add*
IgniteConfiguration setBaselineAutoAdjustEnabled(boolean enabled);
IgniteConfiguration setBaselineAutoAdjustTimeout(long timeoutInMs);
IgniteConfiguration setBaselineAutoAdjustMaxTimeout(long timeoutInMs);
Also, we need to ensure that all nodes would have the same parameters.
And we should be able to survive coordinator left during parameters changes.
-
For IGNITE-8575:
Proposed API format for control.sh:
{{--baseline autoadjust disable}}
{{--baseline autoadjust enable <timeout>}}
was:
From IGNITE-8571:
Now we have only one way to change BLAT - manually update it via console.sh or API.
We need to add the possibility to change it automatically. Adjust to current topology.
So, I propose 3 new parameters which would be responsible to tune this feature.
1. Flag autoAdjustEnabled - true/false. Easy. Manual baseline control or auto adjusting baseline.
2. autoAdjustTimeout - time which we would wait after the actual topology change. But it would be reset if new discovery event happened. (node join/exit).
3. autoAdjustMaxTimeout - time which we would wait from the first dicovery event in the chain. If we achieved it than we would change BLAT right away (no matter were another node join/exit happened or not).
We need to change API next way:
1. org.apache.ignite.IgniteCluster
*Add*
isBaselineAutoAdjustEnabled()
setBaselineAutoAdjustEnabled(boolean enabled);
setBaselineAutoAdjustTimeout(long timeoutInMs);
setBaselineAutoAdjustMaxTimeout(long timeoutInMs);
2. org.apache.ignite.configuration.IgniteConfiguration
*Add*
IgniteConfiguration setBaselineAutoAdjustEnabled(boolean enabled);
IgniteConfiguration setBaselineAutoAdjustTimeout(long timeoutInMs);
IgniteConfiguration setBaselineAutoAdjustMaxTimeout(long timeoutInMs);
Also, we need to ensure that all nodes would have the same parameters.
And we should be able to survive coordinator left during parameters changes.
-
For IGNITE-8575:
Proposed API format for control.sh:
{{--baseline autoadjust disable}}
{{--baseline autoadjust enable <softTimeout> <hardTimeout>}}
> Document Baseline auto-adjust feature
> -------------------------------------
>
> Key: IGNITE-10945
> URL: https://issues.apache.org/jira/browse/IGNITE-10945
> Project: Ignite
> Issue Type: Task
> Components: documentation
> Reporter: Ivan Bessonov
> Assignee: Artem Budnikov
> Priority: Major
> Labels: IEP-4, Phase-2
> Fix For: 2.8
>
>
> From IGNITE-8571:
> Now we have only one way to change BLAT - manually update it via console.sh or API.
> We need to add the possibility to change it automatically. Adjust to current topology.
> So, I propose 3 new parameters which would be responsible to tune this feature.
> 1. Flag autoAdjustEnabled - true/false. Easy. Manual baseline control or auto adjusting baseline.
> 2. autoAdjustTimeout - time which we would wait after the actual topology change. But it would be reset if new discovery event happened. (node join/exit).
> 3. autoAdjustMaxTimeout - time which we would wait from the first dicovery event in the chain. If we achieved it than we would change BLAT right away (no matter were another node join/exit happened or not).
> We need to change API next way:
> 1. org.apache.ignite.IgniteCluster
> *Add*
> isBaselineAutoAdjustEnabled()
> setBaselineAutoAdjustEnabled(boolean enabled);
> setBaselineAutoAdjustTimeout(long timeoutInMs);
> setBaselineAutoAdjustMaxTimeout(long timeoutInMs);
> 2. org.apache.ignite.configuration.IgniteConfiguration
> *Add*
> IgniteConfiguration setBaselineAutoAdjustEnabled(boolean enabled);
> IgniteConfiguration setBaselineAutoAdjustTimeout(long timeoutInMs);
> IgniteConfiguration setBaselineAutoAdjustMaxTimeout(long timeoutInMs);
> Also, we need to ensure that all nodes would have the same parameters.
> And we should be able to survive coordinator left during parameters changes.
> -
> For IGNITE-8575:
> Proposed API format for control.sh:
> {{--baseline autoadjust disable}}
> {{--baseline autoadjust enable <timeout>}}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)