[
https://issues.apache.org/jira/browse/IGNITE-8911?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16724989#comment-16724989
]
Dmitriy Govorukhin commented on IGNITE-8911:
--------------------------------------------
[~EdShangGG] I reviewed your changes and I have several comments:
# You use initDirLock as GridStripedReadWriteLock, but use only writeLock,
why? Write lock is not striped, it is exclusive.
# I see you create a new exception WaitTopologyException for await topology
version, I think will be better create restart future and return
IgniteCacheRestartingException. The user will decide to wait or not wait for
this future, in current implementation we always will be wait for
WaitTopologyException and if topology changed again we will be waiting again.
> While cache is restarting it's possible to start new cache with this name
> -------------------------------------------------------------------------
>
> Key: IGNITE-8911
> URL: https://issues.apache.org/jira/browse/IGNITE-8911
> Project: Ignite
> Issue Type: Bug
> Reporter: Eduard Shangareev
> Assignee: Eduard Shangareev
> Priority: Major
> Fix For: 2.8
>
>
> We have the state "restarting" for caches when we certainly now that these
> caches would start at some moment in future. But we could start new cache
> with the same name.
> Plus, NPE is throwing when we try to get proxy for such caches (in
> "restarting" state).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
