[Gluster-devel] GlusterFlow 0.0.1 :)

Justin Clift jclift at redhat.com
Wed May 8 17:24:27 UTC 2013


Hey Jeff,

*Finally* got all of the pieces of the GlusterFlow thing I've been
working on functional (end to end):

  https://github.com/justinclift/glusterflow

Written in Python 2.7 and using Django + PostgreSQL database.  There's
a screenshot on the front page of the repo. :)

It's still super super basic, but the (first lot of) heavy lifting is
done.  

It uses a Glupy translator to send JSON messages to a central
collector server for every single fop done.  (Yeah, lots of data, and
likely to need optimisation)

The web front shows what's been happening graphically, and (will):

  + allow the user to interactively select time ranges (historical
analysis)
  + drill down into details of what's been used when (which fops,
    which files, ~microsecond accuracy)
  + anything else useful we can think of with this kind of data set
    like "most used files in time range XYZ" "server with most traffic
    in time range ABC" etc.

The graphical stuff is all rendered client side using javascript,
thus giving interactive capability.  Mousing over the columns already
provides stats for the given column's time range, etc.

This could be pretty nifty if we get it right. :)

Note - Anyone wanting to pitch in is welcome to.  Pull request's very
much accepted. :)  For "put it on the Forge" thoughts... later.

+ Justin

--
Open Source and Standards @ Red Hat

twitter.com/realjustinclift





More information about the Gluster-devel mailing list