You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Vijay (Created) (JIRA)" <ji...@apache.org> on 2011/10/19 17:53:10 UTC
[jira] [Created] (CASSANDRA-3383) Make Row cache reject wide rows
which are more than x columns
Make Row cache reject wide rows which are more than x columns
-------------------------------------------------------------
Key: CASSANDRA-3383
URL: https://issues.apache.org/jira/browse/CASSANDRA-3383
Project: Cassandra
Issue Type: Improvement
Components: Core
Affects Versions: 1.1
Environment: Linux with Sun JDK
Reporter: Vijay
Assignee: Vijay
Priority: Minor
The problem with the row cache is it has to cache everything in memory, this works great for small rows but not for the wide rows. Even of off-heap cache has to read everything into the heap eventually causing GC overhead.
It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-3383) Make Row cache reject wide rows
which are more than x columns
Posted by "Jonathan Ellis (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-3383?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13130935#comment-13130935 ]
Jonathan Ellis commented on CASSANDRA-3383:
-------------------------------------------
The idea with the query cache is that you'd be saying "here's a query that I do a lot, that does a cacheable subset of the row." So... no, I don't see where a separate upper bound concept really makes sense (since slice queries already have a count limit, and name queries are obviously limited by nature).
> Make Row cache reject wide rows which are more than x columns
> -------------------------------------------------------------
>
> Key: CASSANDRA-3383
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3383
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.1
> Environment: Linux with Sun JDK
> Reporter: Vijay
> Assignee: Vijay
> Priority: Minor
>
> The problem with the row cache is it has to cache the whole row into the memory, this works great for small rows but not for the wide rows. Even off-heap cache has to read everything into the heap eventually causing GC overhead (If your query is to get few columns).
> It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Closed] (CASSANDRA-3383) Make Row cache reject wide rows
which are more than x columns
Posted by "Vijay (Closed) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-3383?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Vijay closed CASSANDRA-3383.
----------------------------
> Make Row cache reject wide rows which are more than x columns
> -------------------------------------------------------------
>
> Key: CASSANDRA-3383
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3383
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.1
> Environment: Linux with Sun JDK
> Reporter: Vijay
> Assignee: Vijay
> Priority: Minor
>
> The problem with the row cache is it has to cache the whole row into the memory, this works great for small rows but not for the wide rows. Even off-heap cache has to read everything into the heap eventually causing GC overhead (If your query is to get few columns).
> It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-3383) Make Row cache reject wide rows
which are more than x columns
Posted by "Vijay (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-3383?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13130926#comment-13130926 ]
Vijay commented on CASSANDRA-3383:
----------------------------------
Hi Jonathan,
I was actually talking about 2. even with the "query cache" will it make sense to add the upper bound? so that we dont waste resources doing copies?
> Make Row cache reject wide rows which are more than x columns
> -------------------------------------------------------------
>
> Key: CASSANDRA-3383
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3383
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.1
> Environment: Linux with Sun JDK
> Reporter: Vijay
> Assignee: Vijay
> Priority: Minor
>
> The problem with the row cache is it has to cache the whole row into the memory, this works great for small rows but not for the wide rows. Even off-heap cache has to read everything into the heap eventually causing GC overhead (If your query is to get few columns).
> It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (CASSANDRA-3383) Make Row cache reject wide rows
which are more than x columns
Posted by "Vijay (Resolved) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-3383?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Vijay resolved CASSANDRA-3383.
------------------------------
Resolution: Not A Problem
Cool, Makes sense... thanks!
> Make Row cache reject wide rows which are more than x columns
> -------------------------------------------------------------
>
> Key: CASSANDRA-3383
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3383
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.1
> Environment: Linux with Sun JDK
> Reporter: Vijay
> Assignee: Vijay
> Priority: Minor
>
> The problem with the row cache is it has to cache the whole row into the memory, this works great for small rows but not for the wide rows. Even off-heap cache has to read everything into the heap eventually causing GC overhead (If your query is to get few columns).
> It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Commented] (CASSANDRA-3383) Make Row cache reject wide rows
which are more than x columns
Posted by "Jonathan Ellis (Commented) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-3383?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13130768#comment-13130768 ]
Jonathan Ellis commented on CASSANDRA-3383:
-------------------------------------------
My reaction would be that
1) if you really have different data sets involved, the "right" solution would be to split those into different CFs
2) but if you have the same data set where some outlier rows are much larger than others then a "query cache" like CASSANDRA-1956 might be a good general purpose solution
> Make Row cache reject wide rows which are more than x columns
> -------------------------------------------------------------
>
> Key: CASSANDRA-3383
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3383
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.1
> Environment: Linux with Sun JDK
> Reporter: Vijay
> Assignee: Vijay
> Priority: Minor
>
> The problem with the row cache is it has to cache the whole row into the memory, this works great for small rows but not for the wide rows. Even off-heap cache has to read everything into the heap eventually causing GC overhead (If your query is to get few columns).
> It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (CASSANDRA-3383) Make Row cache reject wide rows
which are more than x columns
Posted by "Vijay (Updated) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/CASSANDRA-3383?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Vijay updated CASSANDRA-3383:
-----------------------------
Description:
The problem with the row cache is it has to cache the whole row into the memory, this works great for small rows but not for the wide rows. Even off-heap cache has to read everything into the heap eventually causing GC overhead (If your query is to get few columns).
It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
was:
The problem with the row cache is it has to cache everything in memory, this works great for small rows but not for the wide rows. Even of off-heap cache has to read everything into the heap eventually causing GC overhead.
It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
> Make Row cache reject wide rows which are more than x columns
> -------------------------------------------------------------
>
> Key: CASSANDRA-3383
> URL: https://issues.apache.org/jira/browse/CASSANDRA-3383
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Affects Versions: 1.1
> Environment: Linux with Sun JDK
> Reporter: Vijay
> Assignee: Vijay
> Priority: Minor
>
> The problem with the row cache is it has to cache the whole row into the memory, this works great for small rows but not for the wide rows. Even off-heap cache has to read everything into the heap eventually causing GC overhead (If your query is to get few columns).
> It will be nice to reject the rows which are bigger than x columns (This can be a KS setting).... this will help cases where we have mixed load of small and wide column datasets in the same cf.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira