You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ignite.apache.org by "Anton Vinogradov (JIRA)" <ji...@apache.org> on 2015/08/06 09:31:04 UTC
[jira] [Assigned] (IGNITE-614) Race in preloading/unswapping
[ https://issues.apache.org/jira/browse/IGNITE-614?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Anton Vinogradov reassigned IGNITE-614:
---------------------------------------
Assignee: Anton Vinogradov (was: Semen Boikov)
> Race in preloading/unswapping
> -----------------------------
>
> Key: IGNITE-614
> URL: https://issues.apache.org/jira/browse/IGNITE-614
> Project: Ignite
> Issue Type: Bug
> Components: cache
> Reporter: Semen Boikov
> Assignee: Anton Vinogradov
> Priority: Critical
> Fix For: sprint-7
>
>
> During unswap entry is removed from swap then unswap listener is notified, and this logic in preload supply worker:
> - register unswap listener
> - iterater over heap
> - iterate over swap
> - remove unswap listener
> - iterate over entries recorded by unswap listener
> There is a race: entry is removed from unswap, listener is not notified yet, at this moment supply worker removes unswap listener.
> These tests fail from time to time because entry can be missed during preloading:
> GridCacheSwapPreloadSelfTest.testSwapPartitionedMultithreaded
> GridCacheSwapPreloadSelfTest.testSwapReplicatedMultithreaded
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)