[
https://issues.apache.org/jira/browse/HBASE-15314?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15187759#comment-15187759
]
Daniel Pol commented on HBASE-15314:
------------------------------------
My favorite use case is when you have the hottest table that you want to cache
completely and make sure you get the best performance for it (and the table
doesn't fit into RAM). Right now its a matter of adding software raid on top of
multiple SSDs to achieve that. I would like to remove the software raid
overhead by doing the parallelism in Hbase. I agree it's not easy, mostly
because the Hbase blocksize is not really fixed. So you have to end up adding
logic to handle that.
Funny you mentioned the buckets of fixed size. I'm thinking about filling
another JIRA related to a lot of wasted space in bucketcache because of that.
When you have a small bucketcache that's not an issue, but when you get to a
few TiB and you end up with half the space unused but allocated it becomes a
serious issue.
> Allow more than one backing file in bucketcache
> -----------------------------------------------
>
> Key: HBASE-15314
> URL: https://issues.apache.org/jira/browse/HBASE-15314
> Project: HBase
> Issue Type: Sub-task
> Components: BucketCache
> Reporter: stack
> Assignee: Amal Joshy
> Attachments: HBASE-15314.patch
>
>
> Allow bucketcache use more than just one backing file: e.g. chassis has more
> than one SSD in it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
