You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Mikhail Petrov (Jira)" <ji...@apache.org> on 2022/01/19 14:15:00 UTC
[jira] [Updated] (IGNITE-16331) [Extensions] Add implementation of pluggable topology validator.
[ https://issues.apache.org/jira/browse/IGNITE-16331?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Mikhail Petrov updated IGNITE-16331:
------------------------------------
Description:
It is needed to add an implementation of the Ignite plugin to solve the problem of data inconsistency in case of a cluster split into several independent parts due to network problems.
This problem can be solved by providing the guarantee that
after cluster segmentation, no more than one segment can process write requests to all caches.
It is proposed to
1. reuse Ignite topology validation mechanism to block cache writes
2. automatically apply topology validation logic for all caches through CacheTopologyValidatorProvider
3. rely on remaining Ignite baseline nodes in the topology to determine
what cluster part can continue to work if any
4. automatically switch state of the cluster part to READ-ONLY if it is considered to stop processing of write requests
5. Changing cluster part to active automatically restore writes availability to all caches
> [Extensions] Add implementation of pluggable topology validator.
> ----------------------------------------------------------------
>
> Key: IGNITE-16331
> URL: https://issues.apache.org/jira/browse/IGNITE-16331
> Project: Ignite
> Issue Type: Improvement
> Reporter: Mikhail Petrov
> Assignee: Mikhail Petrov
> Priority: Major
>
> It is needed to add an implementation of the Ignite plugin to solve the problem of data inconsistency in case of a cluster split into several independent parts due to network problems.
> This problem can be solved by providing the guarantee that
> after cluster segmentation, no more than one segment can process write requests to all caches.
> It is proposed to
> 1. reuse Ignite topology validation mechanism to block cache writes
> 2. automatically apply topology validation logic for all caches through CacheTopologyValidatorProvider
> 3. rely on remaining Ignite baseline nodes in the topology to determine
> what cluster part can continue to work if any
> 4. automatically switch state of the cluster part to READ-ONLY if it is considered to stop processing of write requests
> 5. Changing cluster part to active automatically restore writes availability to all caches
--
This message was sent by Atlassian Jira
(v8.20.1#820001)