Question on btm-2.2 (NioJournal)?

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Question on btm-2.2 (NioJournal)?

MVS, Ratnakar
Hi,
 
We are looking for alternatives to existing DiskJournal (of btm-2.1.4, as we’ve performance issues with it) and came across this JIRA item - https://jira.codehaus.org/browse/BTM-107.
 
With the above JIRA, we identified that someone contributed alternative to DiskJournal. For an experimental purpose, we’ve integrated that contribution NioJournal into 2.1.4 distro and able to perform some junit tests. Later we’ve seen that Bitronix has branch code base for v2.2 that contains this NioJournal, but when actually using v2.2 jar (build locally with minimal modifications) we’ve been seeing same kind of performance issues with what we are facing with DiskJournal.
 
Though this NioJournal is not included in official releases of btm’s, but still we would like to know some insights on this to get better understanding. Could you please help us with reasons. This information would really help us on usage of NioJournal.
 
Thanks in advance.
 
Best Regards,
 
 
<img src="rtfimage://" width="353" height="353">
Ratnakar MVS
Software Engineer (II)
NCR Corporation 
office:
+91.40.30891041
mobile: +91.89788.84881
    www.ncr.com       [hidden email]
 
 
 
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Question on btm-2.2 (NioJournal)?

Brett Wooldridge-2
Hi Ratnakar,

I'd like to suggest that you try the 3.0-SNAPSHOT.  The master branch is at https://github.com/bitronix/btm.

Among the changes in 3.0 is:

  • High-performance transaction log journaling using a new design that allows concurrent appenders through a write-reservation model. The new journal is 3-14x faster than BTM 2.1.
The journaling changes are not the NIO writer, but are a major rewrite of the 2.x journaling that dramatically increases performance and eliminates most if not all of the blocking that occurred in the journaling subsystem.

Give it a try, and let us know how it performs for you.  Additionally, if you have lots of long-running transactions, you might try increasing the log file sizes (but not too much, as that can affect performance as well).

-Brett

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Question on btm-2.2 (NioJournal)?

MVS, Ratnakar

Thank you Brett for your prompt response. We would give a try with 3.0 version.

 

Thanks & Regards,

 

Ratnakar MVS
Software Engineer (II)

NCR Corporation 
office:
+91.40.30891041

mobile: +91.89788.84881
    www.ncr.com       [hidden email]

 

From: Brett Wooldridge [mailto:[hidden email]]
Sent: Wednesday, February 18, 2015 8:51 PM
To: [hidden email]
Subject: Re: [btm-user] Question on btm-2.2 (NioJournal)?

 

Hi Ratnakar,

 

I'd like to suggest that you try the 3.0-SNAPSHOT.  The master branch is at https://github.com/bitronix/btm.

 

Among the changes in 3.0 is:

 

  • High-performance transaction log journaling using a new design that allows concurrent appenders through a write-reservation model. The new journal is 3-14x faster than BTM 2.1.

The journaling changes are not the NIO writer, but are a major rewrite of the 2.x journaling that dramatically increases performance and eliminates most if not all of the blocking that occurred in the journaling subsystem.

 

Give it a try, and let us know how it performs for you.  Additionally, if you have lots of long-running transactions, you might try increasing the log file sizes (but not too much, as that can affect performance as well).

 

-Brett

 


image001.png (61K) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Question on btm-2.2 (NioJournal)?

MVS, Ratnakar
In reply to this post by Brett Wooldridge-2

Hi Brett,

 

Thanks once again J

 

We’ve included this SNAPSHOT in our local tests, and the results are looking good. We are also planning to do couple of performance tests on the same.

 

BTW, Are there any plans to release 3.0 version in the near future?

 

Thanks & Regards,

 

Ratnakar MVS
Software Engineer (II)

NCR Corporation 
office:
+91.40.30891041

mobile: +91.89788.84881
    www.ncr.com       [hidden email]

 

From: MVS, Ratnakar
Sent: Wednesday, February 18, 2015 8:53 PM
To: '[hidden email]'
Cc: [hidden email]
Subject: RE: [btm-user] Question on btm-2.2 (NioJournal)?

 

Thank you Brett for your prompt response. We would give a try with 3.0 version.

 

Thanks & Regards,

 

Ratnakar MVS
Software Engineer (II)

NCR Corporation 
office:
+91.40.30891041

mobile: +91.89788.84881
    www.ncr.com       [hidden email]

 

From: Brett Wooldridge [[hidden email]]
Sent: Wednesday, February 18, 2015 8:51 PM
To: [hidden email]
Subject: Re: [btm-user] Question on btm-2.2 (NioJournal)?

 

Hi Ratnakar,

 

I'd like to suggest that you try the 3.0-SNAPSHOT.  The master branch is at https://github.com/bitronix/btm.

 

Among the changes in 3.0 is:

 

  • High-performance transaction log journaling using a new design that allows concurrent appenders through a write-reservation model. The new journal is 3-14x faster than BTM 2.1.

The journaling changes are not the NIO writer, but are a major rewrite of the 2.x journaling that dramatically increases performance and eliminates most if not all of the blocking that occurred in the journaling subsystem.

 

Give it a try, and let us know how it performs for you.  Additionally, if you have lots of long-running transactions, you might try increasing the log file sizes (but not too much, as that can affect performance as well).

 

-Brett

 


image001.png (61K) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

RE: Question on btm-2.2 (NioJournal)?

MVS, Ratnakar
In reply to this post by Brett Wooldridge-2

Hi Brett,

 

Thank you. We’ve included this v3.0 version of Bitronix into our performance test bed and the outcome looks promising.

 

Since it is not yet released to the public audiences, we are reluctant to use this SNAPSHOT in the production environment. I’ve one final query – Could you please let us know “Are there any further updates that needs to come to DiskJournal subsystem?”

 

Thank you very much for your support.

 

Best Regards,

 

Ratnakar MVS
Software Engineer (II)

NCR Corporation 
office:
+91.40.30891041

mobile: +91.89788.84881
    www.ncr.com       [hidden email]

 

From: Brett Wooldridge [mailto:[hidden email]]
Sent: Wednesday, February 18, 2015 8:51 PM
To: [hidden email]
Subject: Re: [btm-user] Question on btm-2.2 (NioJournal)?

 

Hi Ratnakar,

 

I'd like to suggest that you try the 3.0-SNAPSHOT.  The master branch is at https://github.com/bitronix/btm.

 

Among the changes in 3.0 is:

 

  • High-performance transaction log journaling using a new design that allows concurrent appenders through a write-reservation model. The new journal is 3-14x faster than BTM 2.1.

The journaling changes are not the NIO writer, but are a major rewrite of the 2.x journaling that dramatically increases performance and eliminates most if not all of the blocking that occurred in the journaling subsystem.

 

Give it a try, and let us know how it performs for you.  Additionally, if you have lots of long-running transactions, you might try increasing the log file sizes (but not too much, as that can affect performance as well).

 

-Brett

 


image001.png (61K) Download Attachment
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Question on btm-2.2 (NioJournal)?

Ludovic Orban-2
Hi Ratnakar,

As far as I can tell, BTM 3.0.0 is of production quality as it is today. There might be a few minor things to change here and there, but nothing fundamental like the disk journal need any kind of update.

Unfortunately, the project leads (ie: Brett and myself) don't have time to work on BTM anymore. I personally would welcome some motivated helping hands as I'm sure Brett would too.

--
Ludovic

On Sat, Feb 21, 2015 at 4:30 AM, MVS, Ratnakar <[hidden email]> wrote:

Hi Brett,

 

Thank you. We’ve included this v3.0 version of Bitronix into our performance test bed and the outcome looks promising.

 

Since it is not yet released to the public audiences, we are reluctant to use this SNAPSHOT in the production environment. I’ve one final query – Could you please let us know “Are there any further updates that needs to come to DiskJournal subsystem?”

 

Thank you very much for your support.

 

Best Regards,

 

Ratnakar MVS
Software Engineer (II)

NCR Corporation 
office:
<a href="tel:%2B91.40.30891041" value="+914030891041" target="_blank">+91.40.30891041

mobile: <a href="tel:%2B91.89788.84881" value="+918978884881" target="_blank">+91.89788.84881
    www.ncr.com       [hidden email]

 

From: Brett Wooldridge [mailto:[hidden email]]
Sent: Wednesday, February 18, 2015 8:51 PM
To: [hidden email]
Subject: Re: [btm-user] Question on btm-2.2 (NioJournal)?

 

Hi Ratnakar,

 

I'd like to suggest that you try the 3.0-SNAPSHOT.  The master branch is at https://github.com/bitronix/btm.

 

Among the changes in 3.0 is:

 

  • High-performance transaction log journaling using a new design that allows concurrent appenders through a write-reservation model. The new journal is 3-14x faster than BTM 2.1.

The journaling changes are not the NIO writer, but are a major rewrite of the 2.x journaling that dramatically increases performance and eliminates most if not all of the blocking that occurred in the journaling subsystem.

 

Give it a try, and let us know how it performs for you.  Additionally, if you have lots of long-running transactions, you might try increasing the log file sizes (but not too much, as that can affect performance as well).

 

-Brett

 


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Question on btm-2.2 (NioJournal)?

Brett Wooldridge-2
Ludovic,

I'm going to try to publish a 3.0 release following the release process on codehaus.org:


If I run into any difficulties, I'll let you know.


On Mon, Feb 23, 2015 at 9:48 PM, Ludovic Orban <[hidden email]> wrote:
Hi Ratnakar,

As far as I can tell, BTM 3.0.0 is of production quality as it is today. There might be a few minor things to change here and there, but nothing fundamental like the disk journal need any kind of update.

Unfortunately, the project leads (ie: Brett and myself) don't have time to work on BTM anymore. I personally would welcome some motivated helping hands as I'm sure Brett would too.

--
Ludovic

On Sat, Feb 21, 2015 at 4:30 AM, MVS, Ratnakar <[hidden email]> wrote:

Hi Brett,

 

Thank you. We’ve included this v3.0 version of Bitronix into our performance test bed and the outcome looks promising.

 

Since it is not yet released to the public audiences, we are reluctant to use this SNAPSHOT in the production environment. I’ve one final query – Could you please let us know “Are there any further updates that needs to come to DiskJournal subsystem?”

 

Thank you very much for your support.

 

Best Regards,

 

Ratnakar MVS
Software Engineer (II)

NCR Corporation 
office:
<a href="tel:%2B91.40.30891041" value="+914030891041" target="_blank">+91.40.30891041

mobile: <a href="tel:%2B91.89788.84881" value="+918978884881" target="_blank">+91.89788.84881
    www.ncr.com       [hidden email]

 

From: Brett Wooldridge [mailto:[hidden email]]
Sent: Wednesday, February 18, 2015 8:51 PM
To: [hidden email]
Subject: Re: [btm-user] Question on btm-2.2 (NioJournal)?

 

Hi Ratnakar,

 

I'd like to suggest that you try the 3.0-SNAPSHOT.  The master branch is at https://github.com/bitronix/btm.

 

Among the changes in 3.0 is:

 

  • High-performance transaction log journaling using a new design that allows concurrent appenders through a write-reservation model. The new journal is 3-14x faster than BTM 2.1.

The journaling changes are not the NIO writer, but are a major rewrite of the 2.x journaling that dramatically increases performance and eliminates most if not all of the blocking that occurred in the journaling subsystem.

 

Give it a try, and let us know how it performs for you.  Additionally, if you have lots of long-running transactions, you might try increasing the log file sizes (but not too much, as that can affect performance as well).

 

-Brett

 



Loading...