[Gluster-devel] RADOS translator for GlusterFS

[Yehuda Sadeh]

On Mon, May 5, 2014 at 8:21 AM, Jeff Darcy <jdarcy at redhat.com> wrote:
> Now that we're all one big happy family, I've been mulling over
> different ways that the two technology stacks could work together.  One
> idea would be to use some of the GlusterFS upper layers for their
> interface and integration possibilities, but then falling down to RADOS
> instead of GlusterFS's own distribution and replication.  I must
> emphasize that I don't necessarily think this is The Right Way for
> anything real, but I think it's an important experiment just to see what
> the problems are and how well it performs.  So here's what I'm thinking.
>
> For the Ceph folks, I'll describe just a tiny bit of how GlusterFS
> works.  The core concept in GlusterFS is a "translator" which accepts
> file system requests and generates file system requests in exactly the
> same form.  This allows them to be stacked in arbitrary orders, moved
> back and forth across the server/client divide, etc.  There are several
> broad classes of translators:
>
> * Some, such as FUSE or GFAPI, inject new requests into the translator
>   stack.
>
> * Some, such as "posix", satisfy requests by calling a server-local FS.
>
> * The "client" and "server" translators together get requests from one
>   machine to another.
>
> * Some translators *route* requests (one in to one of several out).
>
> * Some translators *fan out* requests (one in to all of several out).
>
> * Most are one in, one out, to add e.g. locks or caching etc.
>
> Of particular interest here are the DHT (routing/distribution) and AFR
> (fan-out/replication) translators, which mirror functionality in RADOS.
> My idea is to cut out everything from these on below, in favor of a
> translator based on librados instead.  How this works is pretty obvious
> for file data - just read and write to RADOS objects instead of to
> files.  It's a bit less obvious for metadata, especially directory

Sorry if I'm missing something obvious, but how are reads / writes
actually done? Do you keep an open file descriptor and work on that
(e.g., are there open() / close() operations), or are operations don't
require any state? With RADOS it's the latter case, so we don't
provide certain guarantees and there are no file-state operations
(like open(), close(), lock(), etc.). Anything like that needs to be
implemented on top of it.

> entries.  One really simple idea is to store metadata as data, in some
> format defined by the translator itself, and have it handle the
> read/modify/write for adding/deleting entries and such.  That would be

Maybe integrate it with the mds (which by itself stores metadata as
data and does all the relevant work)?

> enough to get some basic performance tests done.  A slightly more
> sophisticated idea might be to use OSD class methods to do the
> read/modify/write, but I don't know much about that mechanism so I'm not
> sure that's even feasible.

I don't see why it wouldn't work. The rados gateway does things
similarly for handling the bucket index.

>
> This is not something I'm going to be working on as part of my main job,
> but I'd like to get the experiment started in some of my "spare" time.
> Is there anyone else interested in collaborating, or are there any other
> obvious ideas I'm missing?
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo at vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html