[
https://issues.apache.org/jira/browse/IGNITE-8855?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ivan Bessonov updated IGNITE-8855:
----------------------------------
Description:
After fixing client nodes hangs in IGNITE-8657 another issue was found out:
when EXCHANGE_HISTORY is significantly smaller than number of clients they
interfere with each other on initial exchange and make reconnect attempts over
and over again.
To avoid this random delay (maybe exponential) should be added for clients
before making new reconnect attempt.
was:
After fixing client nodes hangs in IGNITE-8567 another issue was found out:
when EXCHANGE_HISTORY is significantly smaller than number of clients they
interfere with each other on initial exchange and make reconnect attempts over
and over again.
To avoid this random delay (maybe exponential) should be added for clients
before making new reconnect attempt.
> Client nodes make a lot of attempts to join topology if EXCHANGE_HISTORY is
> significantly smaller than number of clients
> ------------------------------------------------------------------------------------------------------------------------
>
> Key: IGNITE-8855
> URL: https://issues.apache.org/jira/browse/IGNITE-8855
> Project: Ignite
> Issue Type: Bug
> Reporter: Sergey Chugunov
> Assignee: Ivan Bessonov
> Priority: Major
>
> After fixing client nodes hangs in IGNITE-8657 another issue was found out:
> when EXCHANGE_HISTORY is significantly smaller than number of clients they
> interfere with each other on initial exchange and make reconnect attempts
> over and over again.
> To avoid this random delay (maybe exponential) should be added for clients
> before making new reconnect attempt.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
