HI,
The prime focus of GlusterFS is to accumulate distributed storages and make
them look like one big one.At the wiki documentation(which in fact is the
only available docs I could find) , they have clearly stated that they are
totally against file striping(though still they provide a translator for
that).In my thinking file striping is the only way to maximixe throughput of
a filesystem as seen by the processess.
It would be better to test and compare b/w using IOR using MPI-IO.
p.s. Do you have some more documents regarding GlusterFS.If yes then please
mail them to me
On 21 Sep 2007 16:05:22 -0000, C Rap <krrapp at rediffmail.com>
wrote:>
> is gluster any good as HPC parallel file system? looks more like a user
> level file system. i''m in the process of evaluating PFS and was
wondering if
> gluster stands any chance against lustre. there don;t seem to be any HPC
> benchmarks listed anywhere.
> sorry for barging like this. kr.
>
>
>
> _______________________________________________
> Lustre-discuss mailing list
> Lustre-discuss at clusterfs.com
> https://mail.clusterfs.com/mailman/listinfo/lustre-discuss
>
>
--
Regards--
Rishi Pathak
National PARAM Supercomputing Facility
Center for Development of Advanced Computing(C-DAC)
Pune University Campus,Ganesh Khind Road
Pune-Maharastra
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mail.clusterfs.com/pipermail/lustre-discuss/attachments/20070922/1b384037/attachment.html