Sensei still alive?

[Casey Sybrandy]

Hello,

Is Sensei still actively supported? We are considering using this for a project at work, but one person on the team heard that it may not be actively developed anymore. It looks to be a good solution to several problems that we're working on, so we're hoping that it's still supported.

Thanks.

[John Wang]

It is absolutely.

We have development teams at LinkedIn, Twitter, Jive and Xiaomi working on this.

Most of the development discussions happen on the developer mailing list and the code activities are on each individual's branches. That's why maybe you are not seeing much activity.

Feel free to post to the mailing list for any questions you may have.

-John

--
You received this message because you are subscribed to the Google Groups "Sensei" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sensei-searc...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Casey Sybrandy]

Cool, thanks!

[Chris Holmes]

I am curious how active this is because the web site refers to supporting applications at LinkedIn that have been retired. Additionally, refers to items being earmarked for completion of integration in "early 2012" such as dynamic schema support.

I am interested in leveraging Sensei; however, not if the project is no longer being updated or under development.

Has dynamic schema been implemented?
Do you leverage HDFS for storage of larger sized JSON objects?
Is there a security model around Sensei?
Why is the open source web site so out of date?

[John Wang]

Hi Chris:

     Yes, Sensei is still being actively worked on. In fact we are in the middle of a release upgrade for Lucene 4.4.

      Since none of us are at LinkedIn anymore for a while, we don't know what is the status there. Last I know the developers at LinkedIn has a fork for the project to address specific needs for their applications. But we have not seen much contributions from that team for a while.

      Instead, most recent contributions have been from Xiaomi and Jive software.

      Most discussions have been on a private developer's list which gives the false impression that Sensei is becoming stale. I am beginning to think we should post to the public mailing list instead.

      Having said that, a substantial number of developers have moved to Twitter. We have been so busy that we have not been able to contribute compare to when we were at LinkedIn. So I would agree progress has been slowed down. But we hope to do more and contribute some of the ideas used by Twitter and apply that to Sensei.

      Specifically about your questions:

1) Dynamic schema: We are not prioritizing this. Our focus has been more on lucene 4.x upgrade and improving performance of the inner core.

2) HDFS storage: HDFS is defintely used for offline indexes. It is an important part of Sensei. As for large sized JSONs, HDFS can definitely handle this. In practice, most companies are storing more compact forms, such as thrift, protocol buffer, avro etc. due to the verbosity of json.

3) What do you mean by security model?

4) OS website: do you mean senseidb.com? I don't think it is out of date. Which parts do you mean specifically?

Cheers

-John