You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@ignite.apache.org by "Valentin Kulichenko (JIRA)" <ji...@apache.org> on 2016/04/12 01:12:25 UTC

[jira] [Created] (IGNITE-2975) Load balancer does not survive client reconnect

Valentin Kulichenko created IGNITE-2975:
-------------------------------------------

             Summary: Load balancer does not survive client reconnect
                 Key: IGNITE-2975
                 URL: https://issues.apache.org/jira/browse/IGNITE-2975
             Project: Ignite
          Issue Type: Bug
          Components: compute
    Affects Versions: 1.5.0.final
            Reporter: Valentin Kulichenko
             Fix For: 1.6


{{RoundRobinGlobalLoadBalancer}} has a snapshot of nodes ({{nodeList}}) which is updated on each discovery event. But when the client reconnects after the full topology restart, the snapshot is not updated and it contains only nodes that already failed. {{EVT_CLIENT_NODE_DISCONNECTED}} and {{EVT_CLIENT_NODE_RECONNECTED}} should be properly handled.

Other load balancing SPIs should be also checked.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)