[
https://issues.apache.org/jira/browse/IGNITE-446?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14364642#comment-14364642
]
Yakov Zhdanov commented on IGNITE-446:
--------------------------------------
If https://issues.apache.org/jira/browse/IGNITE-489 is implemented, then
validator can examine if all the data is visible or not (in case some
partitions are lost without being fully preloaded).
> Implement cache topology validator
> ----------------------------------
>
> Key: IGNITE-446
> URL: https://issues.apache.org/jira/browse/IGNITE-446
> Project: Ignite
> Issue Type: Task
> Reporter: Yakov Zhdanov
> Assignee: Anton Vinogradov
> Fix For: sprint-3
>
>
> I think we should provide user with ability to validate cache topology before
> each cache operation.
> Now user cannot validate the topology of cache update, but this seems to be a
> good feature. I can validate task topology and throw exception on MAP step,
> but I cannot do that on cache put.
> I suggest introducing special topology validator. Whenever cache topology
> changes - it is passed to validator. Validator can check nodes number and
> presence of particular (if user have servers with special role in topology),
> this also may be used to validate segment local node operates in. If cache
> topology is not valid each cache update throws exception until cache topology
> does not become valid back.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
