[ 
https://issues.apache.org/jira/browse/IGNITE-9111?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16561917#comment-16561917
 ] 

ASF GitHub Bot commented on IGNITE-9111:
----------------------------------------

GitHub user Jokser opened a pull request:

    https://github.com/apache/ignite/pull/4455

    IGNITE-9111 Do not wait for deactivation cluster transition

    

You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/gridgain/apache-ignite ignite-9111

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/ignite/pull/4455.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #4455
    
----
commit 2a7f6501e06c9c5c2ab4db521ef4f37419d753bc
Author: Pavel Kovalenko <jokserfn@...>
Date:   2018-07-30T11:44:32Z

    IGNITE-9111 Do not wait for deactivation.

commit 247b8babb49baecc00e91d541dd9f3cecda5675e
Author: Pavel Kovalenko <jokserfn@...>
Date:   2018-07-30T13:23:03Z

    IGNITE-9111 Added test.

----


> Do not wait for deactivation in GridClusterStateProcessor#publicApiActiveState
> ------------------------------------------------------------------------------
>
>                 Key: IGNITE-9111
>                 URL: https://issues.apache.org/jira/browse/IGNITE-9111
>             Project: Ignite
>          Issue Type: Improvement
>          Components: cache
>    Affects Versions: 2.4, 2.5
>            Reporter: Pavel Kovalenko
>            Assignee: Pavel Kovalenko
>            Priority: Critical
>             Fix For: 2.7
>
>
> Currently, we wait for activation/deactivation future when check state of the 
> cluster. But when deactivation is in progress it doesn't make sense to wait 
> for it, because after the successful wait we will throw an exception that 
> cluster is not active. Synchronous waiting for deactivation future may lead 
> to deadlocks if operation obtains some locks before checking cluster state.
> As the solution, we should check and wait only for activation futures. In 
> case of in-progress deactivation, we should fail fast and return "false" from 
> publicApiActiveState method.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to