[ https://issues.apache.org/jira/browse/HDFS-6988?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14145610#comment-14145610 ]
Colin Patrick McCabe commented on HDFS-6988: -------------------------------------------- I don't understand why we can't simply have this be a single number, the number of free bytes we should keep around in the ramdisk. Perhaps I'm missing something, but I don't see why percentage-based eviction is useful. Why would I want the number of bytes that I keep free at all times to double when my ramdisk size goes from 4g to 8g, for example? I'm also concerned that this is not integrated with HDFS-4949 cache pools. Since we will have clusters running both, we need some way of sharing memory between the two. Can you make this pluggable? That way, we could later write an eviction threshold policy which integrates with cache pools (perhaps by making the write cache another cache pool). > Add configurable limit for percentage-based eviction threshold > -------------------------------------------------------------- > > Key: HDFS-6988 > URL: https://issues.apache.org/jira/browse/HDFS-6988 > Project: Hadoop HDFS > Issue Type: Sub-task > Components: datanode > Affects Versions: HDFS-6581 > Reporter: Arpit Agarwal > Assignee: Arpit Agarwal > Fix For: HDFS-6581 > > Attachments: HDFS-6988.01.patch, HDFS-6988.02.patch > > > Per feedback from [~cmccabe] on HDFS-6930, we can make the eviction > thresholds configurable. The hard-coded thresholds may not be appropriate for > very large RAM disks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)