Re: disperse volume and its relation to xlators/cluster/ec

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Dear  Xavi,

  How  is  the  data fragmented ? what if  a file  size  is  more than  the fragment size? Can a fragment have data corresponding to 2 or more files ? 
  How  the  data is  accessed  basically ( file open ?) ?  What  is  a  frame ?
  I have  searched a lot  of  documents about  glusterFS. None  of these  are  helpful  to understand  the  glusterFS in detail (from the perspective of  writing  a  new translator),
  Jeff Darcy's tutorial  was helpful though.
   Perhaps, am I missing something?

  Can you  point to some good stuffs ?

Best Regards
JK

On Wed, Feb 17, 2016 at 4:24 PM, jayakrishnan mm <jayakrishnan.mm@xxxxxxxxx> wrote:
Dear Xavi,

Thanks  a lot. That clarified my confusion.

Best Regards
JK

On Wed, Feb 17, 2016 at 3:35 PM, Xavier Hernandez <xhernandez@xxxxxxxxxx> wrote:
Hi jayakrishnan,

the old implementation was not finally accepted to be included in main glusterfs tree, so it was rewritten in what now is known as ec.

disperse is an alias for ec. They are the same. The algorithm implemented is Reed-Solomon. It's really similar to ida (we could say that ida is a specific implementation of a Reed-Solomon code).

Almost all basic functionality in old code is implemented in new ec in some way, though not using the same approaches. Anyway the latest implementation is way more stable and maintained, while the old code has been abandoned and should not be used.

Regards,

Xavi


On 17/02/16 05:12, jayakrishnan mm wrote:
Dear  Xavier,

I am trying to understand the disperse  translator  and its  usage.

 From
https://lists.gnu.org/archive/html/gluster-devel/2014-01/txttzloLYIJOh.txt
, I see  there  are  four  components namely gfsys,dfc,ida  and  heal
which needs  to  be  compiled  with GlusterFS main source code.
I am  confused  this with xlator/cluster/ec.   What is the relationship
between this  and  the disperse volume components  mentioned  earlier ?
  Is   xlator/cluster/ec  is  erasure coding (ec) translator
implementation  in glusterFS ? If yes, which algorithm it implements ? (
I know  disperse  volume implements IDA)
Pls. help

Best Regards
JK


_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users

[Index of Archives]     [Gluster Development]     [Linux Filesytems Development]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux