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

Posted to issues@ignite.apache.org by "Yakov Zhdanov (JIRA)" <ji...@apache.org> on 2015/12/15 16:28:46 UTC

[jira] [Created] (IGNITE-2170) DHT preloader: response to partitions request may be sent bypassing waits for ops to finish

Yakov Zhdanov created IGNITE-2170:
-------------------------------------

             Summary: DHT preloader: response to partitions request may be sent bypassing waits for ops to finish
                 Key: IGNITE-2170
                 URL: https://issues.apache.org/jira/browse/IGNITE-2170
             Project: Ignite
          Issue Type: Task
            Reporter: Yakov Zhdanov


When topology changes, new oldest node may request partitions from other nodes. Other nodes reply immediately, disregarding locks and ongoing TXs. They should reply only after ongoing stuff is finished.
Please check if this is correct.
See org.gridgain.grid.kernal.processors.cache.distributed.dht.preloader.GridDhtPreloader#processSinglePartitionRequest
We reply directly. It seems that we should go through exchange future and reply only when initFut of exch fut is done.

(GG-7284)




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