cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jeff Jirsa (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (CASSANDRA-13486) Upstream Power arch specific code
Date Wed, 03 May 2017 23:56:04 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-13486?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15995871#comment-15995871
] 

Jeff Jirsa edited comment on CASSANDRA-13486 at 5/3/17 11:55 PM:
-----------------------------------------------------------------

{quote}
As this is a feature based on a proprietary technology and the implementation seems to play
nicely with the plugable row cache API, I wonder if it wouldn't make more sense to bundle
this as separate jar for interested users.
{quote}

Perhaps, though where would we identify it? How would users find it? Edit: we have GCE and
AWS snitches in tree, why would this be any different? It's a public extension of a public
interface usable by a portion of the userbase. As long as it's using the public API, it seems
reasonable to be in tree? 

{quote}
The license of capiblock.jar is Apache License, Version 2.0.
{quote}

We typically include the license in [lib/licenses|https://github.com/apache/cassandra/tree/trunk/lib/licenses]
to record this fact. 

{quote}
About the plugin approach, my impression is that Cassandra has not been so pluggable, and
a few exceptions are row cache, metrics, etc. (aside from a couple of unsuccessful proposals
like pluggable storage; correct me if I am wrong.) Is it a direction of Cassandra to make
it more pluggable if it makes sense and to keep and support the plugin APIs? 
{quote}

Cassandra has many pluggable components including (but not limited to):

- Replication strategy
- Seed provider
- Compaction strategy
- Endpoint Snitch
- Secondary indexes
- Authentication / Authorization providers

What Cassandra doesn't have is a good way to promote or market or support third party pluggable
implementations of these components if they're not in tree, and we don't have a way to test
/ verify them if they are in tree (if someone wanted to change the interface for some future
version of Cassandra, none of the current committers have that hardware to verify that it
works as intended). 




was (Author: jjirsa):
{quote}
As this is a feature based on a proprietary technology and the implementation seems to play
nicely with the plugable row cache API, I wonder if it wouldn't make more sense to bundle
this as separate jar for interested users.
{quote}

Perhaps, though where would we identify it? How would users find it? 

{quote}
The license of capiblock.jar is Apache License, Version 2.0.
{quote}

We typically include the license in [lib/licenses|https://github.com/apache/cassandra/tree/trunk/lib/licenses]
to record this fact. 

{quote}
About the plugin approach, my impression is that Cassandra has not been so pluggable, and
a few exceptions are row cache, metrics, etc. (aside from a couple of unsuccessful proposals
like pluggable storage; correct me if I am wrong.) Is it a direction of Cassandra to make
it more pluggable if it makes sense and to keep and support the plugin APIs? 
{quote}

Cassandra has many pluggable components including (but not limited to):

- Replication strategy
- Seed provider
- Compaction strategy
- Endpoint Snitch
- Secondary indexes
- Authentication / Authorization providers

What Cassandra doesn't have is a good way to promote or market or support third party pluggable
implementations of these components if they're not in tree, and we don't have a way to test
/ verify them if they are in tree (if someone wanted to change the interface for some future
version of Cassandra, none of the current committers have that hardware to verify that it
works as intended). 



> Upstream Power arch specific code
> ---------------------------------
>
>                 Key: CASSANDRA-13486
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-13486
>             Project: Cassandra
>          Issue Type: Task
>          Components: Core
>            Reporter: Amitkumar Ghatwal
>         Attachments: CAPI_RowCache_JIRA_Description.doc
>
>
> Hi All/ [~snazy], [~jasobrown] , [~ReiOdaira] ,
> As was suggested to my previous JIRA ticket : https://issues.apache.org/jira/browse/CASSANDRA-13345
to create a separate development branch ( for power -ppc64le ) by forking github/apache/cassandra
. 
> I have created my own development branch : https://github.com/ghatwala/cassandra/tree/ppc64le-CAPI/trunk
and pushed in power arch specific features of CAPI ( Row Cache ). 
> Please refer PR :https://github.com/apache/cassandra/pull/108 to see all the code changes
for CAPI-Row Cache . Please kindly review the same.
> Also please refer to the description document on CAPI-Row Cache.
> Regards,
> Amit



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org


Mime
View raw message