[
https://issues.apache.org/jira/browse/IGNITE-24974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Roman Puchkovskiy updated IGNITE-24974:
---------------------------------------
Description:
The problem is caused by zone locks in PartitionReplicaLifecycleManager; these
locks block threads of TableManager-io thread pool; the same pool is used to
unlock the locks.
We could switch the locks to async mode: when a lock cannot be acquired, a lock
request would simply be enqueued and satisfied later (causing completion of the
corresponding future) instead of blocking the thread.
> Pool starvation when stopping zone in colocation mode
> -----------------------------------------------------
>
> Key: IGNITE-24974
> URL: https://issues.apache.org/jira/browse/IGNITE-24974
> Project: Ignite
> Issue Type: Bug
> Reporter: Roman Puchkovskiy
> Assignee: Roman Puchkovskiy
> Priority: Major
> Labels: ignite-3
>
> The problem is caused by zone locks in PartitionReplicaLifecycleManager;
> these locks block threads of TableManager-io thread pool; the same pool is
> used to unlock the locks.
> We could switch the locks to async mode: when a lock cannot be acquired, a
> lock request would simply be enqueued and satisfied later (causing completion
> of the corresponding future) instead of blocking the thread.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
