[
https://issues.apache.org/jira/browse/IGNITE-16782?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sergey Uttsel updated IGNITE-16782:
-----------------------------------
Description:
During development, I discovered that we have
*org.apache.ignite.internal.table.distributed.storage.InternalTableImpl#partitionMap*
and
*org.apache.ignite.internal.table.distributed.storage.InternalTableImpl#updatePartMapMux*
that do not always go together, which can cause problems when changing this
map in parallel and reading it.
It seems we can eliminate some of the concurrency issues using a
ConcurrentHashMap instead of Int2ObjectMap. But to eliminate all of the
concurrency issues need to more complex investigation.
was:During development, I discovered that we have
*org.apache.ignite.internal.table.distributed.storage.InternalTableImpl#partitionMap*
and
*org.apache.ignite.internal.table.distributed.storage.InternalTableImpl#updatePartMapMux*
that do not always go together, which can cause problems when changing this
map in parallel and reading it.
> Revision of work with InternalTableImpl#partitionMap
> ----------------------------------------------------
>
> Key: IGNITE-16782
> URL: https://issues.apache.org/jira/browse/IGNITE-16782
> Project: Ignite
> Issue Type: Bug
> Reporter: Kirill Tkalenko
> Assignee: Vladislav Pyatkov
> Priority: Major
> Labels: ignite-3
> Fix For: 3.0.0-alpha5
>
>
> During development, I discovered that we have
> *org.apache.ignite.internal.table.distributed.storage.InternalTableImpl#partitionMap*
> and
> *org.apache.ignite.internal.table.distributed.storage.InternalTableImpl#updatePartMapMux*
> that do not always go together, which can cause problems when changing this
> map in parallel and reading it.
>
> It seems we can eliminate some of the concurrency issues using a
> ConcurrentHashMap instead of Int2ObjectMap. But to eliminate all of the
> concurrency issues need to more complex investigation.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
