[jira] [Commented] (IGNITE-24373) Memory leak in ZonePartitionReplicaListener

["Aleksandr Polovtsev (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IGNITE-24373?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17922807#comment-17922807 ] 

Aleksandr Polovtsev commented on IGNITE-24373:
----------------------------------------------

I would like to add two additional points:
1. Under colocation feature, table removal doesn't work, because safe time sync is not implemented, which blocks some futures related to storage removal. So, it is very hard to write a test for this ticket without substantial hacks until IGNITE-22620 is implemented.
2. The same issue will exist for partition-level Raft listeners after IGNITE-22459 will be implemented, it is expected that it will be fixed as well here. 

> Memory leak in ZonePartitionReplicaListener
> -------------------------------------------
>
>                 Key: IGNITE-24373
>                 URL: https://issues.apache.org/jira/browse/IGNITE-24373
>             Project: Ignite
>          Issue Type: Bug
>            Reporter: Aleksandr Polovtsev
>            Assignee: Aleksandr Polovtsev
>            Priority: Minor
>
> {{ZonePartitionReplicaListener}} contains a {{replicas}} map that contains table-level partition listeners. These listeners are never removed from this map, which means that when a table is dropped, partition listeners from this table will stay in memory until the whole zone partition is stopped.



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