[ 
https://issues.apache.org/jira/browse/IGNITE-8544?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Pavel Kovalenko updated IGNITE-8544:
------------------------------------
    Description: 
After exchange is done, we're using initial exchange version to determine 
topology version on what rebalance should be finished and save it. After 
rebalance finishing we check current topology version and saved version and if 
they are equal, we enable WAL, own partitions and do checkpoint. In other case 
we do nothing, because of topology change. 
In case of exchanges merge we're saving old topology version (before merge) and 
it leads to ignoring logic of enabling WAL and etc, because check on topology 
version will be always false-negative.


  was:
After exchange is done, we're using initial exchange version to determine 
topology version on what rebalance should be finished and save it. After 
rebalance finishing we check current topology version and saved version and if 
they are equal, we enable WAL, own partitions and do checkpoint. In other case 
we do nothing, because of topology change. 
In case of exchanges merge we're saving old topology version (before merge) and 
it leads to ignoring logic of enabling WAL and etc, because check on topology 
version will be always false-positive.



> WAL disabling during rebalance mechanism uses wrong topology version in case 
> of exchanges merge
> -----------------------------------------------------------------------------------------------
>
>                 Key: IGNITE-8544
>                 URL: https://issues.apache.org/jira/browse/IGNITE-8544
>             Project: Ignite
>          Issue Type: Bug
>          Components: cache
>    Affects Versions: 2.5
>            Reporter: Pavel Kovalenko
>            Assignee: Pavel Kovalenko
>            Priority: Critical
>             Fix For: 2.6
>
>
> After exchange is done, we're using initial exchange version to determine 
> topology version on what rebalance should be finished and save it. After 
> rebalance finishing we check current topology version and saved version and 
> if they are equal, we enable WAL, own partitions and do checkpoint. In other 
> case we do nothing, because of topology change. 
> In case of exchanges merge we're saving old topology version (before merge) 
> and it leads to ignoring logic of enabling WAL and etc, because check on 
> topology version will be always false-negative.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to