UK Ruby on Rails, Exalead, AWS, Consultancy - InfiniDB, Infobright and MonetDB - Day 1: InfiniDB Comments

Comment on 'InfiniDB, Infobright and MonetDB - Day 1: InfiniDB' by John Weber, Calpont

2010-06-21T22:54:01Z

Hi Chris,

Just thought I would drop by one last time to let you know that we are almost to the point of stamping the "Final" label on our latest release, version 1.5. With this release, we have added quite a bit of functionality over and above the original release earlier this year. Most notably with 1.5 you'll find support for deployment on the Windows operating system and support for sub queries.

We'd love to hear your thoughts if you've found time to give InfiniDB a spin. As always, let us know if you any questions.

John http://infinidb.org http://calpont.com

Comment on 'InfiniDB, Infobright and MonetDB - Day 1: InfiniDB' by Chris Anderton

2010-02-16T16:54:15Z

Hi John

Thanks for dropping by - yes - i noticed you'd gone final and have already downloaded it in order to give it another try - i hope to do so in the next week or two so will likely post again on the results...

Thanks

Comment on 'InfiniDB, Infobright and MonetDB - Day 1: InfiniDB' by John Weber, Calpont

2010-02-12T16:00:58Z

Chris,

I thought I would just take a moment to catch you up on our progress at Calpont with InfiniDB. Since my last post, we have moved the Community Edition forward to a "final" status and, as of last week, we have launched the commercial Enterprise Edition at http://calpont.com. The Enterprise Edition has everything the Community Edition has, but then adds scale-out MPP processing allowing queries to not only be parallelized across cores like the Community Edition, but across servers as well.

A number of other folks have been doing comparison tests of the open source columnar technology similar to this effort. Here's a link to some testing that Vadim at Percona has completed: http://www.mysqlperformanceblog.com/2010/01/07/star-schema-bechmark-infobright-infinidb-and-luciddb/.

Let me know if there is anything we can do to answer questions for you.

John

Comment on 'InfiniDB, Infobright and MonetDB - Day 1: InfiniDB' by Chris Conrey

2009-12-15T02:49:28Z

Looks promising, looking forward to watching more of this project's evolution. Nice find Chris.

Comment on 'InfiniDB, Infobright and MonetDB - Day 1: InfiniDB' by Chris Anderton

2009-11-25T10:32:33Z

Thanks John - we will be taking a more in depth look at InfiniDB over the coming weeks so we'll be sure to keep feeding back - i'm finishing off my whistlestop tour and we can then decide how they can be of use to our clients of the short, medium and long term.

Comment on 'InfiniDB, Infobright and MonetDB - Day 1: InfiniDB' by John Weber, Calpont

2009-11-24T23:33:20Z

Thank you for taking the time to download and try the InfiniDB software. You correctly mentioned in your blog post that the software is currently in an Alpha state, one that we hope the Community, like you, will help us advance quickly. We depend on this type of feedback, so we're grateful for learning about your experience so far.

You'll be happy to know that we are currently working on extending support for VARCHAR from 255 to 8000, so hopefully this will help with whatever string data you are attemting to store. That should be available to the Community here within the next few weeks. Also, I want to thank you for using the Launchpad bug tracking system to enter the above bug. We've been able to duplicate your issue and that has been assigned to a developer. This fix should also be available shortly.

Further on the issue that requires a reboot, it's possible that there is also a runtime setting that needs to be adjusted to match the amount of memory available on the server that could be contributing to your experience. For table join operations, we have a setting within the calpont.xml configuration file called UmMaxMemorySmallSide. If it is unchanged on your system, it is likely currently set to 4G. If you have 4GB or less of memory available on the machine that you are running on, there is a chance that you could be running out of system memory for a given operation and thus ending up in the state mentioned above. A safe setting would be in the neighborhood of 25% of available memory on the system, so for a server with 4GB, try setting this value to 1G. In addition, there is also a NumBlocksPct setting that potentially will need adjustment as well to keep in balance with the UmMaxMemorySmallSide setting. Both of these settings and more are covered in our free Performance and Tuning Guide, available on infinidb.org for registered users. If you have not taken the time to do so, I would strongly recommend taking a moment to register and download that document...I suspect that it will prove to be a useful asset for you.

As you continue your evaluation, please continue to use Launchpad or the InfiniDB Forums to log your issues. We'll get on them just as quickly as we can once they're reported.

We certainly want you and everyone out there to have a positive experience with the software so we're here to help. I would be happy to get on the phone with you or would also be happy to work through your issues on this blog, on the community forums or on Launchpad. Just let us know what works for you.