[jira] [Resolved] (HDDS-3267) Replace ContainerCache in BlockUtils by LoadingCache

["Istvan Fajth (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HDDS-3267?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Istvan Fajth resolved HDDS-3267.
--------------------------------
    Resolution: Won't Fix

> Replace ContainerCache in BlockUtils by LoadingCache
> ----------------------------------------------------
>
>                 Key: HDDS-3267
>                 URL: https://issues.apache.org/jira/browse/HDDS-3267
>             Project: Hadoop Distributed Data Store
>          Issue Type: Improvement
>            Reporter: Isa Hekmatizadeh
>            Assignee: Isa Hekmatizadeh
>            Priority: Minor
>              Labels: pull-request-available
>          Time Spent: 10m
>  Remaining Estimate: 0h
>
> As discussed in [here|https://github.com/apache/hadoop-ozone/pull/705] current version of ContainerCache is just used by BlockUtils and has several architectural issues. for example:
>  * It uses a ReentrantLock which could be replaced by synchronized methods
>  * It should maintain a referenceCount for each DBHandler
>  * It extends LRUMap while it would be better to hide it by the composition and not expose LRUMap related methods.
> As [~pifta] suggests, we could replace all ContainerCache functionality by using Guava LoadingCache.
> This new LoadingCache could be configured to evict by size, by this configuration the functionality would be slightly different as it may evict DBHandlers while they are in use (referenceCount>0) but we can configure it to use reference base eviction based on CacheBuilder.weakValues() 
> I want to open this discussion here instead of Github so I created this ticket.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: ozone-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: ozone-issues-help@hadoop.apache.org