[
https://issues.apache.org/jira/browse/IGNITE-8905?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16535717#comment-16535717
]
Maxim Muzafarov commented on IGNITE-8905:
-----------------------------------------
[~dpavlov], [~sergey-chugunov],
Folks,
As we have stable execution timeouts for some test suites (e.g. currently at
Activate\Deactivate cluster suite) TeamCity Run:All would not show as that we
on the safe side with this fix.
Thus I think it's a good practice to have for PR a special reproducer for
described case.
I've reviewed test cases added by IGNITE-8657
{{testClientReconnectOnExchangeHistoryExhaustion}},
{{testClientInForceServerModeStopsOnExchangeHistoryExhaustion}} and they are
not cover this issue.
What do you think?
> Incorrect assertion in GridDhtPartitionsExchangeFuture
> ------------------------------------------------------
>
> Key: IGNITE-8905
> URL: https://issues.apache.org/jira/browse/IGNITE-8905
> Project: Ignite
> Issue Type: Bug
> Reporter: Sergey Chugunov
> Assignee: Sergey Chugunov
> Priority: Major
> Fix For: 2.7
>
> Original Estimate: 2h
> Time Spent: 1h 1m
> Remaining Estimate: 59m
>
> Assertion was added as part IGNITE-8657 into GridDhtPartitionsExchangeFuture
> which is correct only in situation when client has to reconnect due to too
> short EXCHANGE_HISTORY.
> Exceptions from other situations like not being able to acquire file lock are
> also passed to client node in FullMessage.
> This assertion should be removed and check should be introduced instead: if
> this exception is intended to be thrown on current client node, we should do
> this, otherwise old program flow should be executed.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
