You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Manish Zope (Created) (JIRA)" <ji...@apache.org> on 2011/12/23 20:18:30 UTC

[jira] [Created] (CASSANDRA-3668) Performance of sstablloader is affected in 1.0.x

Performance of sstablloader is affected in 1.0.x
------------------------------------------------

                 Key: CASSANDRA-3668
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
             Project: Cassandra
          Issue Type: Bug
          Components: API
    Affects Versions: 1.0.7
            Reporter: Manish Zope
             Fix For: 1.0.7


One my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
Due to above reported issue generator problem is solved but performance of the sstableloader is still issue.
Isuue 3589 is marked as duplicate of 3618.Both issues shows resolved status.
But the problem with sstableloader still exists.

So opening other issue so that sstbleloader problem should not go unnoticed.

FYI : We have tested the generator part with the patch given in 3589.Its Working fine.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yuki Morishita updated CASSANDRA-3668:
--------------------------------------

    Attachment:     (was: 0001-Allow-multiple-connection-in-StreamInSession.patch)
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Peter Schuller (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13194562#comment-13194562 ] 

Peter Schuller commented on CASSANDRA-3668:
-------------------------------------------

Ok. Definitely interesting to look into then, because just the fact that we're copying some arrays around is not adequate explanation for how we would bottleneck to the point of only reaching 7 MB/second (that's *nothing*). The indexes/bloom filters sound more plausible, but my understanding was that Manish never observed CPU saturation.

Anyways, I was asking since I was assigned to review the patch. I'll try to do a comparison of my own and see what I can find out.

*Possibly* relatedly I have observed cases where there is clearly something wrong with the way we do I/O. Index sampling on start-up have been observed at 20% CPU utilization with plenty of I/O capacity to spare (this is on 0.8 though). In this case something is clearly outright wrong, and is not about performance bottlenecks as such. It's possible something similar is going on here, which would be very interesting to confirm/deny.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13262708#comment-13262708 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

* I don't think the 'retries' field should be moved from StreamInSession to IncomingStreamReader. The ISR is attached to the streaming of one file, so when the node on the other side does a retry, a new ISR will be created (while the StreamInSession persists until closed), and so retries field will be reseted and we will retry infinitely. But it probably does mean it'll make more sense to have one retry counter per-file (changing StreamInSession.currentFiles to a Map<String, Integer> maybe?)
* Currently, every messages related to the streaming goes through the ISR socket *except* for the message of session failure (in StreamInSession.closeInternal). Could be nice to use this ticket to fix that (but if you prefer moving to another ticket that's ok).
* StreamOutSession.begin() shouldn't assume the queue has at least one element (at least the current code don't expect it)

Smallish remarks:
* MessagingService: Looking at ThreadPoolExecutor code it seems that this is harmless but bumping the executor core pool size would make it exceed the max pool size.  I'd be in favor of changing the executor definition:
  {noformat}
  executor = new DebuggableThreadPoolExecutor(0,
                                              1,
                                              1,
                                              TimeUnit.SECONDS,
                                              new LinkedBlockingQueue<Runnable>(),
                                              new NamedThreadFactory("Streaming to " + to));
  {noformat}
  to
  {noformat}
  executor = new DebuggableThreadPoolExecutor(1,
                                              Integer.MAX_VALUE,
                                              1,
                                              TimeUnit.SECONDS,
                                              new LinkedBlockingQueue<Runnable>(),
                                              new NamedThreadFactory("Streaming to " + to));
  {noformat}
  as it's cleaner imho (it doesn't use the non-documented fact that a corePoolSize of 0 still starts a core thread for instance) and equivalent (since DTPE timeouts core threads).
* IncomingTcpConnection: sets the name when starting streaming but don't unset it. Probably not worth changing the name, the thread name is probably not the best place anyway to log activity (but it's a good idea to set a meaningful name in the first place).
* In FileStreamTask.runMayThrow, instead of logging an error, we could wrap the IOException in a new one with the message. That way we wouldn't have 2 messages in the log for the same error.
* Would feel more coherent to me to move the sending of the SESSION_FINISHED message in ISR (for instance SIS.closeIfFinished could return a boolean if the session is indeed finished (and we would just have to pass the file name to this method instead of the ISR instance)).
* Not this patch fault but in SIS.getIncomingFiles(), adding the currentFiles is useless
* In StreamOutSession.begin() there is a commented debug line (could be removed)
* I think the patch needs one more rebase (after CASSANDRA-4174)

                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13181435#comment-13181435 ] 

Manish Zope commented on CASSANDRA-3668:
----------------------------------------

I am loading around 200 data files with total size of 70 GB.

Please let us know if you need any further details.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Parallel streaming for sstableloader

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis updated CASSANDRA-3668:
--------------------------------------

    Issue Type: Improvement  (was: Bug)
    
> Parallel streaming for sstableloader
> ------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: API
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.3
>
>         Attachments: 3668-1.1.txt, 3668-1.1-v2.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Updated] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Sylvain Lebresne updated CASSANDRA-3668:
----------------------------------------

    Fix Version/s:     (was: 1.1.0)
                   1.1.1
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13194543#comment-13194543 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

bq.  because I don't see where/how it was established that the original problem

To sum up, as as Yuki remarked above, the reason streaming is slower in 1.0 is likely due to 2 aspects (basically because that's the 2 things that changed):
# We switched from NIO transfer to "standard" read and send. Meaning we possibly adds copies here and there during transfer.
# We now build indexes (primary and secondary) and bloom filters sequentially with the streaming while it was previously done in parallel of the streaming.

I don't think we really have established what fraction of the slowdown each is responsible of. But the idea of adding concurrent streaming is to "fix" the second one.

Yuki, you said last patch improves throughput: but how do we stand with this patch compared to 0.8?
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13263750#comment-13263750 ] 

Yuki Morishita commented on CASSANDRA-3668:
-------------------------------------------

Sylvain, thanks for review. I will post updates based on your feedback.

Before that, here is the result of my performance test.
Used 1 node for vanilla C* (0.8.10, 1.1.0, 1.1.0-patched), and 1 node for sstableloader. Generated sstables to stream(8 files, 1.6GB total) using stress tool against v0.8.10 cassandra, and performed sstableloader(streaming throttling is disabled). Same sstable files were used for all tests. Each test ran twice and took average.

|| test case         || time to complete (sec) || avg. bps (MB/S) ||
| v0.8.10            | 52.38(*1) | 53 |
| v1.1.0             | 52.30     | 15 |
| v1.1.0-patched n=1 | 52.61     | 15 |
| v1.1.0-patched n=2 | 32.42     | 27 |
| v1.1.0-patched n=4 | 30.80     | 27 |
| v1.1.0-patched n=8 | 33.88     | 25 |

n value for v1.1.0-patched is number of threads to use.
v0.8.10 sleeps 30 sec to get gossip info, so (*1) is actually 22.38.

It's not as fast as 0.8 with this feature, but can boost loading compared to current versions.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis updated CASSANDRA-3668:
--------------------------------------

    Fix Version/s:     (was: 1.1.1)
                   1.2

I'm also getting a little nervous that this turned out to require deeper changes than we thought at first.  Re-targetting for 1.2.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.2
>
>         Attachments: 3668-1.1-v2.txt, 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Issue Comment Edited] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Issue Comment Edited) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13181435#comment-13181435 ] 

Manish Zope edited comment on CASSANDRA-3668 at 1/6/12 5:06 PM:
----------------------------------------------------------------

We are loading around 200 data files with total size of 70 GB.

Please let us know if you need any further details.

Now we are aiming to load 350 GB size of data with around 1000 data files on daily basis.
With hampered performance of sstableloader after 1.0, its made the things very difficult for us.
                
      was (Author: mzope):
    I am loading around 200 data files with total size of 70 GB.

Please let us know if you need any further details.

Now we are aiming to load 350 GB size of data with around 1000 data files on daily basis.
With hampered performance of sstableloader after 1.0, its made the things very difficult for us.
                  
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13179626#comment-13179626 ] 

Jonathan Ellis commented on CASSANDRA-3668:
-------------------------------------------

But it was also single-threaded in 0.8, right?
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13266560#comment-13266560 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

bq.  I just felt weird to see the thread name of IncomingTCPConn as default (like 'Thread-10') on the other hand OutboundTCPConn gets name like 'WRITE-192.168.1.141'

I agree, and I was good with
{noformat}
super("READ-" + socket.getRemoteSocketAddress());
{noformat}
My "problem" was with
{noformat}
setName("streaming from " + socket.getRemoteSocketAddress());
{noformat}

bq. I also changed the name of variable streamingThreadsPerNode to streamExecutorCoreSize and set its default value to 0, in order to preserve current behavior.

In fact my argument was that using 0 is bad but that was argumented already on the original issue (https://issues.apache.org/jira/browse/CASSANDRA-3494?focusedCommentId=13178349&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13178349) but somehow I forgot to commit my 'cleanup' patch. Anyway, I took the liberty to commit said cleanup patch (so you will need to rebase this one if you don't mind).

bq. I modified to have per session retries, since we are closing session after retries exceed the limit.

Yeah, I suppose the best way to handle that would be a retry-per-file to avoid failing a seesion that may make progress on some files, but that's a bigger refactor than is worth here.


A few additional remarks:
* StreamInSession.isFinished handling is not thread safe, we should use an AtomicBoolean and test-an-set it. However AbstractStreamSession.close() already use the same trick to ensure closeInternal() is called only once, so maybe we could just move the code to add the sstables and secondary indexes to closeInternal. 
* StreamInSession.setCurrentFile() is a no-op and should be removed.

                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yuki Morishita updated CASSANDRA-3668:
--------------------------------------

    Attachment: 3688-reply_before_closing_writer.txt

So far I found one place that can improve throughput.

Right now, node that receive streaming replies finished after finalizing sstable, but that takes some time (few hundred msec / per data file), so sender has to wait that amount of time before transferring next file.

Attached patch modifies this behavior so that node can send reply as soon as receiving whole data file and sender can proceed immediately to send next file.

Manish, could you try this patch and see how it improves? It can be applied on cassandra 1.0.6 as well as the latest cassandra-1.0 branch.

I'm still investigating other improvements especially buffering. If I find anything, I will post them here.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis updated CASSANDRA-3668:
--------------------------------------

    Reviewer: scode  (was: jbellis)
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Issue Comment Edited] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Issue Comment Edited) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13179591#comment-13179591 ] 

Manish Zope edited comment on CASSANDRA-3668 at 1/4/12 4:23 PM:
----------------------------------------------------------------

Yuki,

Our test environment consist of cassandra cluster of two nodes and one loader machine.
Readings of the I/O and CPU usage is mentioned in the attached text file.

Specifically we don't see any spikes in cpu usage either on loader or cassandra nodes.
During data load with 0.8.7 setup the average load on the cassandra nodes was higher as compared to 1.0.7.
So I guess cassandra 1.0.7 is not making full use of the resources available on the cassandra nodes.

1.0.7 used here is with updated patch mentioned in issue - 3589.

I might be wrong giving this advice but is it possible to change the UnsortedWriter so that we can generate 
sstables that are already compressed ? This will also help in making loading process even faster as the data 
that need to be streamed will be very less and save the resources required by cassandra nodes for compressing. 


Thanks
Manish
                
      was (Author: mzope):
    Yuki,

Our test environment consist of cassandra cluster of two nodes and one loader machine.
Readings of the I/O and CPU usage is mentioned in the attached text file.

Specifically we don't see any spikes in cpu usage either on loader or cassandra nodes.
During data load with 0.8.7 setup the average load on the cassandra nodes was higher as compared to 1.0.7.
So I guess cassandra 1.0.7 is not making full use of the resources available on the cassandra nodes.

1.0.7 used here is with updated patch mentioned in issue - 3589.

Thanks
Manish
                  
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yuki Morishita updated CASSANDRA-3668:
--------------------------------------

    Attachment: 3668-1.1-v2.txt


Attaching rebased and updated version.

bq. super("READ-" + socket.getRemoteSocketAddress());

is back in ITC.

bq. StreamInSession.isFinished handling is not thread safe, we should use an AtomicBoolean and test-an-set it. However AbstractStreamSession.close() already use the same trick to ensure closeInternal() is called only once, so maybe we could just move the code to add the sstables and secondary indexes to closeInternal.

What I wanted to do here is to prevent sending SESSION_FINISHED message multiple times as well. So instead of put the code in closeInternal, I expose isClosed AtomicBoolean properties as protected and use it in closeIfFiished.

bq. StreamInSession.setCurrentFile() is a no-op and should be removed.

This part of the code was confusing. PendingFiles in 'files' are set beginning of stream session and never get updated unless we swap its content with the one used in IncomingStreamReader. Instead of swapping content, I added updateProgress method to explicitly update progress of pending files.

In addition, there is the problem in StreamOutSession that first stream in session does not contain all pending files in its stream header when stream in parallel. I fixed it by making copy of all pending files before beginning stream. (StreamOutSession#begin)
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 3668-1.1-v2.txt, 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13295130#comment-13295130 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

I believe this is a parallel improvement to CASSANDRA-4297 so it's not made irrelevant by it per-se (I'm talking of Yuki's patch, not of the ticket title).
  
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.2
>
>         Attachments: 3668-1.1-v2.txt, 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13179591#comment-13179591 ] 

Manish Zope commented on CASSANDRA-3668:
----------------------------------------

Yuki,

Our test environment consist of cassandra cluster of two nodes and one loader machine.
Readings of the I/O and CPU usage is mentioned in the attached text file.

Specifically we don't see any spikes in cpu usage either on loader or cassandra nodes.
During data load with 0.8.7 setup the average load on the cassandra nodes was higher as compared to 1.0.7.
So I guess cassandra 1.0.7 is not making full use of the resources available on the cassandra nodes.

1.0.7 used here is with updated patch mentioned in issue - 3589.

Thanks
Manish
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Issue Comment Edited] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Issue Comment Edited) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13181435#comment-13181435 ] 

Manish Zope edited comment on CASSANDRA-3668 at 1/6/12 5:05 PM:
----------------------------------------------------------------

I am loading around 200 data files with total size of 70 GB.

Please let us know if you need any further details.

Now we are aiming to load 350 GB size of data with around 1000 data files on daily basis.
With hampered performance of sstableloader after 1.0, its made the things very difficult for us.
                
      was (Author: mzope):
    I am loading around 200 data files with total size of 70 GB.

Please let us know if you need any further details.
                  
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177130#comment-13177130 ] 

Manish Zope commented on CASSANDRA-3668:
----------------------------------------

Hi Jonathan,
Sorry for the late reply.

We have created the sstables with 0.8.7 and run the sstableloader from separate machine in our test environment.

On same setup we replaced 0.8.7 with 1.0.x (With updated patch as mentioned in issue -3589).
Time required for loading data of same size took 2x time as compared to that of 0.8.7.





                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Issue Comment Edited] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Issue Comment Edited) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13179437#comment-13179437 ] 

Manish Zope edited comment on CASSANDRA-3668 at 1/4/12 3:37 PM:
----------------------------------------------------------------

Hi Jonathan,

We took some readings for comparison of the sstabloader utility in 0.8.7 and 1.0.7 (With updated patch as mentioned in issue -3589).Please find the description in attached file named as "sstable-loader performance.txt"

Please let us know if you want any further inputs from our side.

Thanks
Manish
                
      was (Author: mzope):
    Hi Jonathan,

We took some readings for comparison of the sstabloader utility in 0.8.7 and 1.0.7.
Please find the description in attached file named as "sstable-loader performance.txt"

Please let us know if you want any further inputs on the same.

Thanks
Manish
                  
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13185061#comment-13185061 ] 

Jonathan Ellis commented on CASSANDRA-3668:
-------------------------------------------

bq. Attached patch modifies this behavior so that node can send reply as soon as receiving whole data file 

Doesn't this mean that if the target restarts after acking the stream, but before finishing the 2ary index builds, that data will never be indexed but we'll have no indication it didn't happen?

If this is the bottleneck I'd rather add a concurrency option to the bulk loader as a workaround.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Parallel streaming for sstableloader

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis updated CASSANDRA-3668:
--------------------------------------

    Affects Version/s:     (was: 1.0.0)
        Fix Version/s:     (was: 1.2.0)
                       1.3
              Summary: Parallel streaming for sstableloader  (was: Performance of sstableloader is affected in 1.0.x)
    
> Parallel streaming for sstableloader
> ------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.3
>
>         Attachments: 3668-1.1.txt, 3668-1.1-v2.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Updated] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yuki Morishita updated CASSANDRA-3668:
--------------------------------------

    Attachment:     (was: 0002-Allow-concurrent-stream-in-StreamOutSession.patch)
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13260680#comment-13260680 ] 

Yuki Morishita commented on CASSANDRA-3668:
-------------------------------------------

Sylvain,

I posted rebased patch here: https://github.com/yukim/cassandra/tree/3668
After I do performance comparison, I will post how it goes along with the formal patch set.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Issue Comment Edited] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (Issue Comment Edited) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13194543#comment-13194543 ] 

Sylvain Lebresne edited comment on CASSANDRA-3668 at 1/27/12 8:41 AM:
----------------------------------------------------------------------

bq.  because I don't see where/how it was established that the original problem was due to a lack of concurrency

To sum up, as as Yuki remarked above, the reason streaming is slower in 1.0 is likely due to 2 aspects (basically because that's the 2 things that changed):
# We switched from NIO transfer to "standard" read and send. Meaning we possibly adds copies here and there during transfer.
# We now build indexes (primary and secondary) and bloom filters sequentially with the streaming while it was previously done in parallel of the streaming.

I don't think we really have established what fraction of the slowdown each is responsible of. But the idea of adding concurrent streaming is to "fix" the second one.

Yuki, you said last patch improves throughput: but how do we stand with this patch compared to 0.8?
                
      was (Author: slebresne):
    bq.  because I don't see where/how it was established that the original problem

To sum up, as as Yuki remarked above, the reason streaming is slower in 1.0 is likely due to 2 aspects (basically because that's the 2 things that changed):
# We switched from NIO transfer to "standard" read and send. Meaning we possibly adds copies here and there during transfer.
# We now build indexes (primary and secondary) and bloom filters sequentially with the streaming while it was previously done in parallel of the streaming.

I don't think we really have established what fraction of the slowdown each is responsible of. But the idea of adding concurrent streaming is to "fix" the second one.

Yuki, you said last patch improves throughput: but how do we stand with this patch compared to 0.8?
                  
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13185069#comment-13185069 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

Isn't CASSANDRA-3494 that concurrency option?
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13180910#comment-13180910 ] 

Yuki Morishita commented on CASSANDRA-3668:
-------------------------------------------

Right...

In 0.8, we use java.nio to transfer files. We switched to normal java.io in 1.0 because of compression (both for sstable and internode streaming) and index building during streaming.
So basically transferring files is faster in 0.8, though 0.8 needs to build indexes and BFs after transferring file.

Manish, how many files are you loading? And could you tell me the size of those files?
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177549#comment-13177549 ] 

Manish Zope commented on CASSANDRA-3668:
----------------------------------------

Yuki/Jonathan,

Thanks for the advice.Let me try and will share output of the test with you shortly.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13175561#comment-13175561 ] 

Jonathan Ellis commented on CASSANDRA-3668:
-------------------------------------------

And to clarify: we're still talking about compared to 0.8.7 right?
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13218046#comment-13218046 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

@Yukim Would you mind rebasing?

Also, you said getting better throughput with this patch, but could you give a few more details on that (like numbers and what kind of tests you performed).

I moved this to 1.1.1. This modify streaming and I'm not confortable doing that on 1.1.0 at this point of the freeze. 
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13185087#comment-13185087 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

You're right.

Probably wouldn't be too hard to allow the per-targer executor to use more than one thread (and modify streaming accordingly to it's able to give to each thread a different file). And that would be streaming generic actually, so all the better for bootstrap and such.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Assigned] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Assigned) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis reassigned CASSANDRA-3668:
-----------------------------------------

    Assignee: Yuki Morishita

Can you tell us how to reproduce?  What kind of degradation are you seeing?
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13295131#comment-13295131 ] 

Yuki Morishita commented on CASSANDRA-3668:
-------------------------------------------

Yes. CASSANDRA-4297 improves CPU utility but overall throughput is still about the same.
What I'm trying to do for this ticket is to enable parallel streaming through sstableloader.
I guess I should update the title.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.2
>
>         Attachments: 3668-1.1-v2.txt, 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Sylvain Lebresne (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13267348#comment-13267348 ] 

Sylvain Lebresne commented on CASSANDRA-3668:
---------------------------------------------

bq. What I wanted to do here is to prevent sending SESSION_FINISHED message multiple times as well.

Yes, but we must equally avoid to add multiple times the sstables to the CFS, which can happen if you reuse isClosed that way because when checking its value in closeIfFinished we don't set it at the same time atomically. What I'm suggesting is to move all the code that adds sstables and build 2ndary indexes to closeInternal (only in the case where it's successful) since close guarantee it is called only once, and then, we could make AbstractStreamSession.close return true if it did perform the closing, and false if it was already closed and return that value in closeIfFinished to avoid multiple send of SESSION_FINISHED.

bq. In addition, there is the problem in StreamOutSession that first stream in session does not contain all pending files in its stream header when stream in parallel.

Right, which means we have an additional problem: MessagingService.stream() is non-blocking, so it's possible another file is streamed before the first one has been started and thus before the session has been created on the receiving node. The problem is, the correct way to fix that would probably be to change streaming to 1) send the stream header and wait for the other node to acknowledge the reception and 2) start streaming files. But such change pretty much imply changing the target of this to 1.2.

Another option could be to wait for the first file to be streamed before starting other ones, but that would diminish the interest of this patch. Though I'm starting to wonder if it's really reasonable to push this to a "stable" release anyway.

                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 3668-1.1-v2.txt, 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Peter Schuller (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13194459#comment-13194459 ] 

Peter Schuller commented on CASSANDRA-3668:
-------------------------------------------

So I am a bit confused by the history of this ticket. Manish, as far as I can tell, your observations were that the streaming of a single file is slower than 1.0, correct? I.e., you were watching iostats and CPU usage while a large file was being streamed?

There is definitely merit to supporting concurrent streaming for a given source/target, I'm not arguing against that. But I am wondering if I am misunderstanding something, because I don't see where/how it was established that the original problem was due to a lack of concurrency. Certainly if there is a huge amount of sstables the latency will cause lower average throughput; but was it established that this is what's happening? With 350 gigs, even with 1000 sstables, the average sstable should be large enough that it seems unlikely that the major contributor to poor performance is the waiting period between streams.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177432#comment-13177432 ] 

Yuki Morishita commented on CASSANDRA-3668:
-------------------------------------------

Manish,

sstableLloader loads sstables by streaming data files to the cluster, and from 1.0, CASSANDRA-3080 introduced throttling for streaming.
By default, {{stream_throughput_outbound_megabits_per_sec}} is set to 400 Mbps or 50MB/sec.
You can turn off throttling by setting 0 inside cassandra.yaml.

You can observe the behavior of throttling by changing log level to TRACE or DEBUG in conf/log4j-tools.properties, and  look for log with the term "Throttle". (Sample output for throttling is attached in CASSANDRA-3080.)
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Issue Comment Edited] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Issue Comment Edited) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13179437#comment-13179437 ] 

Manish Zope edited comment on CASSANDRA-3668 at 1/4/12 12:33 PM:
-----------------------------------------------------------------

Hi Jonathan,

We took some readings for comparison of the sstabloader utility in 0.8.7 and 1.0.7.
Please find the description in attached file named as "sstable-loader performance.txt"

Please let us know if you want any further inputs on the same.

Thanks
Manish
                
      was (Author: mzope):
    sstable loader readings
                  
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yuki Morishita updated CASSANDRA-3668:
--------------------------------------

    Attachment: 0003-Add-threads-option-to-sstableloader.patch
                0002-Allow-concurrent-stream-in-StreamOutSession.patch
                0001-Allow-multiple-connection-in-StreamInSession.patch

Attached patches add threads option(-t) to sstableloader. The option allows you to configure # of threads per destination.

I made patches for trunk because in 1.0 branch streaming socket is one-to-one to incoming stream session.

I got better throughput with 4 threads, and  observed little impact on target node's cpu and memory.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Manish Zope updated CASSANDRA-3668:
-----------------------------------

    Summary: Performance of sstableloader is affected in 1.0.x  (was: Performance of sstablloader is affected in 1.0.x)
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>             Fix For: 1.0.7
>
>   Original Estimate: 96h
>  Remaining Estimate: 96h
>
> One my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> Due to above reported issue generator problem is solved but performance of the sstableloader is still issue.
> Isuue 3589 is marked as duplicate of 3618.Both issues shows resolved status.
> But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13295099#comment-13295099 ] 

Jonathan Ellis commented on CASSANDRA-3668:
-------------------------------------------

Can this be resolved given CASSANDRA-4297?
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.2
>
>         Attachments: 3668-1.1-v2.txt, 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177616#comment-13177616 ] 

Manish Zope commented on CASSANDRA-3668:
----------------------------------------

Hi Yuki

We tried the sstableloader by setting stream_throughput_outbound_megabits_per_sec to 0 on loader machine 1st and then on cassandra cluster machines.
But unfortunately there is no improvement in streaming.It streams at a avg. speed of 7 MBPS for version 1.0.x which is 35 MBPS for 0.8.7
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13179584#comment-13179584 ] 

Yuki Morishita commented on CASSANDRA-3668:
-------------------------------------------

Manish,

So, does your test environment have just one cassandra node? How many sstable files are you loading?
Do you observe any bottleneck during loading to cassandra node? (cpu spikes, io wait...)
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Manish Zope updated CASSANDRA-3668:
-----------------------------------

    Attachment: sstable-loader performance.txt

sstable loader readings
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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] [Issue Comment Edited] (CASSANDRA-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (Issue Comment Edited) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177432#comment-13177432 ] 

Yuki Morishita edited comment on CASSANDRA-3668 at 12/29/11 10:33 PM:
----------------------------------------------------------------------

Manish,

sstableLloader loads sstables by streaming data files to the cluster, and from 1.0, CASSANDRA-3080 introduced throttling for streaming.
By default, {{stream_throughput_outbound_megabits_per_sec}} is set to 400 Mbps or 50MB/sec.
You can turn off throttling by setting 0 inside cassandra.yaml.

You can observe the behavior of throttling in sstableloader by changing log level to TRACE or DEBUG in conf/log4j-tools.properties, and  look for log with the term "Throttle". (Sample log output for throttling is attached in CASSANDRA-3080.)
                
      was (Author: yukim):
    Manish,

sstableLloader loads sstables by streaming data files to the cluster, and from 1.0, CASSANDRA-3080 introduced throttling for streaming.
By default, {{stream_throughput_outbound_megabits_per_sec}} is set to 400 Mbps or 50MB/sec.
You can turn off throttling by setting 0 inside cassandra.yaml.

You can observe the behavior of throttling by changing log level to TRACE or DEBUG in conf/log4j-tools.properties, and  look for log with the term "Throttle". (Sample output for throttling is attached in CASSANDRA-3080.)
                  
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13185072#comment-13185072 ] 

Jonathan Ellis commented on CASSANDRA-3668:
-------------------------------------------

I don't think so, since that just adds thread-per-target, not multiple-threads-to-single-target.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.8
>
>         Attachments: 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13177220#comment-13177220 ] 

Jonathan Ellis commented on CASSANDRA-3668:
-------------------------------------------

Are you seeing a bottleneck in i/o?  in CPU?

http://spyced.blogspot.com/2010/01/linux-performance-basics.html
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yuki Morishita updated CASSANDRA-3668:
--------------------------------------

    Attachment:     (was: 0003-Add-threads-option-to-sstableloader.patch)
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Yuki Morishita updated CASSANDRA-3668:
--------------------------------------

    Attachment: 3668-1.1.txt

I rebased and attached patch. (also pushed to https://github.com/yukim/cassandra/tree/3668-2).

bq. I don't think the 'retries' field should be moved from StreamInSession to IncomingStreamReader. The ISR is attached to the streaming of one file, so when the node on the other side does a retry, a new ISR will be created (while the StreamInSession persists until closed), and so retries field will be reseted and we will retry infinitely. But it probably does mean it'll make more sense to have one retry counter per-file (changing StreamInSession.currentFiles to a Map<String, Integer> maybe?)

I modified to have per session retries, since we are closing session after retries exceed the limit.

bq. Currently, every messages related to the streaming goes through the ISR socket except for the message of session failure (in StreamInSession.closeInternal).

I moved SESSION_FAILURE to ISR.

bq. StreamOutSession.begin() shouldn't assume the queue has at least one element (at least the current code don't expect it)

Fixed.

bq. MessagingService: Looking at ThreadPoolExecutor code it seems that this is harmless but bumping the executor core pool size would make it exceed the max pool size.

Agree. I changed max pool size to Integer.MAX_VALUE. I also changed the name of variable streamingThreadsPerNode to streamExecutorCoreSize and set its default value to 0, in order to preserve current behavior.

bq. IncomingTcpConnection: sets the name when starting streaming but don't unset it. Probably not worth changing the name, the thread name is probably not the best place anyway to log activity (but it's a good idea to set a meaningful name in the first place).

For now I reverted this change. I just felt weird to see the thread name of IncomingTCPConn as default (like 'Thread-10') on the other hand OutboundTCPConn gets name like 'WRITE-192.168.1.141'.

bq. In FileStreamTask.runMayThrow, instead of logging an error, we could wrap the IOException in a new one with the message. That way we wouldn't have 2 messages in the log for the same error.

You are right. Fixed.

bq. Would feel more coherent to me to move the sending of the SESSION_FINISHED message in ISR (for instance SIS.closeIfFinished could return a boolean if the session is indeed finished (and we would just have to pass the file name to this method instead of the ISR instance)).

Moved to ISR.

bq. Not this patch fault but in SIS.getIncomingFiles(), adding the currentFiles is useless

This was actually necessary because pending files in 'files' field do not get updated. But having separate field for currently streaming files seems redundant, so in this version I modified to replace pending file in 'files' with actually streaming one.

bq. In StreamOutSession.begin() there is a commented debug line (could be removed)

Removed.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>              Labels: streaming
>             Fix For: 1.1.1
>
>         Attachments: 3668-1.1.txt, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Jonathan Ellis (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis updated CASSANDRA-3668:
--------------------------------------

             Reviewer: jbellis
             Priority: Minor  (was: Major)
    Affects Version/s:     (was: 1.0.7)
                       1.0.0
        Fix Version/s:     (was: 1.0.8)
                       1.1
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.0
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>            Priority: Minor
>             Fix For: 1.1
>
>         Attachments: 0001-Allow-multiple-connection-in-StreamInSession.patch, 0002-Allow-concurrent-stream-in-StreamOutSession.patch, 0003-Add-threads-option-to-sstableloader.patch, 3688-reply_before_closing_writer.txt, sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Yuki Morishita (Commented) (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13179598#comment-13179598 ] 

Yuki Morishita commented on CASSANDRA-3668:
-------------------------------------------

CASSANDRA-3494 addresses that streaming is mono threaded, and maybe that is causing the performance problem.
It is planned for 1.1, but I can rebase it for 1.0 branch for testing purpose.
                
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>            Assignee: Yuki Morishita
>             Fix For: 1.0.7
>
>         Attachments: sstable-loader performance.txt
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Manish Zope updated CASSANDRA-3668:
-----------------------------------

    Remaining Estimate: 48h  (was: 96h)
     Original Estimate: 48h  (was: 96h)
    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> Due to above reported issue generator problem is solved but performance of the sstableloader is still issue.
> Isuue 3589 is marked as duplicate of 3618.Both issues shows resolved status.
> But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.

--
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-3668) Performance of sstableloader is affected in 1.0.x

Posted by "Manish Zope (Updated) (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3668?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Manish Zope updated CASSANDRA-3668:
-----------------------------------

    Description: 
One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.

3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.

So opening other issue so that sstbleloader problem should not go unnoticed.

FYI : We have tested the generator part with the patch given in 3589.Its Working fine.

Please let us know if you guys require further inputs from our side.

  was:
One my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
Due to above reported issue generator problem is solved but performance of the sstableloader is still issue.
Isuue 3589 is marked as duplicate of 3618.Both issues shows resolved status.
But the problem with sstableloader still exists.

So opening other issue so that sstbleloader problem should not go unnoticed.

FYI : We have tested the generator part with the patch given in 3589.Its Working fine.

    
> Performance of sstableloader is affected in 1.0.x
> -------------------------------------------------
>
>                 Key: CASSANDRA-3668
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3668
>             Project: Cassandra
>          Issue Type: Bug
>          Components: API
>    Affects Versions: 1.0.7
>            Reporter: Manish Zope
>             Fix For: 1.0.7
>
>   Original Estimate: 48h
>  Remaining Estimate: 48h
>
> One of my colleague had reported the bug regarding the degraded performance of the sstable generator and sstable loader.
> ISSUE :- https://issues.apache.org/jira/browse/CASSANDRA-3589 
> As stated in above issue generator performance is rectified but performance of the sstableloader is still an issue.
> 3589 is marked as duplicate of 3618.Both issues shows resolved status.But the problem with sstableloader still exists.
> So opening other issue so that sstbleloader problem should not go unnoticed.
> FYI : We have tested the generator part with the patch given in 3589.Its Working fine.
> Please let us know if you guys require further inputs from our side.

--
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