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

Posted to issues@ignite.apache.org by "Kirill Tkalenko (Jira)" <ji...@apache.org> on 2023/02/20 06:25:00 UTC

[jira] [Created] (IGNITE-18843) Fix the behavior of MvPartitionStorage#pollForVacuum when trying to delete the same row in parallel

Kirill Tkalenko created IGNITE-18843:
----------------------------------------

             Summary: Fix the behavior of MvPartitionStorage#pollForVacuum when trying to delete the same row in parallel
                 Key: IGNITE-18843
                 URL: https://issues.apache.org/jira/browse/IGNITE-18843
             Project: Ignite
          Issue Type: Bug
            Reporter: Kirill Tkalenko
             Fix For: 3.0.0-beta2


It is necessary to slightly correct the behavior of the *org.apache.ignite.internal.storage.MvPartitionStorage#pollForVacuum* when we are trying to delete the same *RowVersion* in parallel.

At the moment, when trying to delete 2 *RowVersion* in parallel, we will actually delete the first *RowVersion* for one thread, and the second thread will notice this and simply return *null*, and as a result, we will have one *RowVersion* that needs to be deleted.

We need to try to get *RowVersion*s until the queue is empty.

It needs to be fixed for all implementations.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)