This is very interesting work. I have not gone through the details in the doc, but I'm +1 on the idea.
On Fri, Jul 13, 2018 at 9:36 PM Shashikant Banerjee < [email protected]> wrote: > +1. > > Thanks > Shashi > > On 7/14/18, 2:28 AM, "Anu Engineer" <[email protected]> wrote: > > +1, This would be extremely useful for Ozone too. > > Thanks > Anu > > > On 7/13/18, 1:26 PM, "Lokesh Jain" <[email protected]> wrote: > > +1 > > Thanks > Lokesh > > On 14-Jul-2018, at 1:55 AM, Jitendra Pandey < > [email protected]> wrote: > > > > +1 > > > > On 7/13/18, 1:22 PM, "Josh Elser" <[email protected]> wrote: > > > > Hi all, > > > > I'd like to share a document that I've been working on for > many months > > with Jitendra about building a distributed-log "recipe" on > top of Ratis. > > > > > https://docs.google.com/document/d/1Su5py_T5Ytfh9RoTTX2s20KbSJwBHVxbO7ge5ORqbCk/edit?usp=sharing > > > > The primary use case is for Apache HBase, replacing its > HDFS-backed > > write-ahead log construct with something a bit more flexible. > Our hope > > is that by building a distributed log with Ratis, HBase could > observe > > similar levels of performance as we do with HDFS and better > enable our > > flexibility to run on "dynamic" cloud IaaS platforms. > > > > At the bottom of that document, you'll find a high level > breakdown of > > work items for both Ratis as well as HBase. My hope is that > both of > > these will be able to be worked on concurrently. Furthermore, > while > > HBase is a primary "customer" for the work, I hope that you > would find a > > DistributedLog recipe for Ratis (what we've been calling the > > "LogService") generally reusable for the Ratis community. > > > > Please let me know what you think (either on the mailing list > or via > > comments on the google-doc). Thanks! > > > > - Josh > > > > > > > > > > > > -- Mingliang Liu
