[ 
https://issues.apache.org/jira/browse/IGNITE-11416?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Ivan Bessonov updated IGNITE-11416:
-----------------------------------
    Description: 
We need following improvements:
 * do not write the same value twice in a row, this would lead to history 
pollution;
 * add "putAll" functionality on binary level, not in public API yet. This 
would simplify the migration in future;
 * do not use "*HistoryItem" class for everything, this is not conventional;
 * retrieve "dmsVer" from cluster on handshake, this would help to reduce 
joining node DataBag size drastically;
 * add "isEmpty()" or "long getVersion()" method to metastorage, will be 
helpful for components that use it;
 * there has to be an ability to read data on client nodes, maybe write as 
well, not sure yet;
 * implement more optimal in-memory storage for history cache;
 * start in-memory metastorage at proper time, current implementation does it a 
bit too late.

  was:
We need following improvements:
 * do not write the same value twice in a row, this would lead to history 
pollution;
 * add "putAll" functionality on binary level, not in public API yet. This 
would simplify the migration in future;
 * do not use "*HistoryItem" class for everything, this is not conventional;
 * retrieve "dmsVer" from cluster on handshake, this would help to reduce 
joining node DataBag size drastically;
 * add "isEmpty()" or "long getVersion()" method to metastorage, will be 
helpful for components that use it;
 * there has to be an ability to read data on client nodes, maybe write as 
well, not sure yet.


> DistributedMetaStorage improvements
> -----------------------------------
>
>                 Key: IGNITE-11416
>                 URL: https://issues.apache.org/jira/browse/IGNITE-11416
>             Project: Ignite
>          Issue Type: Improvement
>            Reporter: Ivan Bessonov
>            Assignee: Ivan Bessonov
>            Priority: Major
>
> We need following improvements:
>  * do not write the same value twice in a row, this would lead to history 
> pollution;
>  * add "putAll" functionality on binary level, not in public API yet. This 
> would simplify the migration in future;
>  * do not use "*HistoryItem" class for everything, this is not conventional;
>  * retrieve "dmsVer" from cluster on handshake, this would help to reduce 
> joining node DataBag size drastically;
>  * add "isEmpty()" or "long getVersion()" method to metastorage, will be 
> helpful for components that use it;
>  * there has to be an ability to read data on client nodes, maybe write as 
> well, not sure yet;
>  * implement more optimal in-memory storage for history cache;
>  * start in-memory metastorage at proper time, current implementation does it 
> a bit too late.



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

Reply via email to