You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Alexey Zinoviev (Jira)" <ji...@apache.org> on 2019/10/01 14:14:00 UTC

[jira] [Updated] (IGNITE-11163) [ML] Add hart beat tracker to DistributedInfModel

     [ https://issues.apache.org/jira/browse/IGNITE-11163?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Alexey Zinoviev updated IGNITE-11163:
-------------------------------------
    Affects Version/s:     (was: 2.8)
                       2.9

> [ML] Add hart beat tracker to DistributedInfModel
> -------------------------------------------------
>
>                 Key: IGNITE-11163
>                 URL: https://issues.apache.org/jira/browse/IGNITE-11163
>             Project: Ignite
>          Issue Type: Improvement
>          Components: ml
>    Affects Versions: 2.9
>            Reporter: Anton Dmitriev
>            Assignee: Anton Dmitriev
>            Priority: Major
>             Fix For: 2.8
>
>
> Currently we have DistributedInfModel that allocates resources across the cluster and relies on close() method to release them (it implements AutoCloseable). It works well in case client node works correctly. If it doesn't resources might stay allocated and so that we'll get a memory leak.
> To mitigate this risk it's proposed to add hart beat mechanism that will track if client is alive or not from server (service) side.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)