[jira] [Comment Edited] (CASSANDRA-6890) Standardize on a single read path

["Joshua McKenzie (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-6890?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13984870#comment-13984870 ] 

Joshua McKenzie edited comment on CASSANDRA-6890 at 4/29/14 10:02 PM:
----------------------------------------------------------------------

nio2 and FILE_SHARE_DELETE still won't allow us to delete hard-links to memory-mapped files - [reference|http://www.osronline.com/showThread.cfm?link=64732].  While that link refers to files rather than hard-links to files, I tested with nio2 and confirmed it applies to memory-mapped hard-links on Windows as well.

I'll retest perf w/both mixed and cql native; I went in with the expectation that stress defaulted to cql3 at this point given the development focus.  @Pavel - I'll get the perf record results after re-testing on Windows.



was (Author: joshuamckenzie):
nio2 and FILE_SHARE_DELETE still won't allow us to delete hard-links to memory-mapped files - [reference|http://www.osronline.com/showThread.cfm?link=64732].  While that link refers to files in general, I tested with nio2 and confirmed it applies to memory-mapped hard-links on Windows as well.

I'll retest perf w/both mixed and cql native; I went in with the expectation that stress defaulted to cql3 at this point given the development focus.  @Pavel - I'll get the perf record results after re-testing on Windows.


> Standardize on a single read path
> ---------------------------------
>
>                 Key: CASSANDRA-6890
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-6890
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Joshua McKenzie
>            Assignee: Joshua McKenzie
>              Labels: performance
>             Fix For: 3.0
>
>         Attachments: mmap_gc.jpg, mmap_jstat.txt, mmap_perf.txt, nommap_gc.jpg, nommap_jstat.txt
>
>
> Since we actively unmap unreferenced SSTR's and also copy data out of those readers on the read path, the current memory mapped i/o is a lot of complexity for very little payoff.  Clean out the mmapp'ed i/o on the read path.



--
This message was sent by Atlassian JIRA
(v6.2#6252)