[Beowulf] Question on hgh performance, low cost Fileserver

Daniel Pfenniger daniel.pfenniger at obs.unige.ch
Tue Nov 15 10:46:02 PST 2005


What about the low cost original requirement?
Is Terragrid free software to be part of the Beowulf model?
Link?

Thanks,
	Dan

Imran Khan wrote:
> Arvind,
> 
>  
> 
> I think you should try some thing like Terragrid, Terragrid uses a cache
> coherent implementation of iSCSI, to make a standard Linux filesystem
> behave as a parallel filesystem.
> 
>  
> 
> The advantages this brings are:
> 
>  
> 
> 1.                   Standard Linux filesystem, and tools, so no re-training
> 
> 2.                   Small code base of only 40,000 lines of c, which
> means easy to support software
> 
> 3.                   Unlike, PVFS, GFS, Lustre etc, TerraGrid does not
> use a Meta Data Controller, so scales linearly.
> 
> 4.                   TerraGrid is the only CFS solution with a 24X7
> resilient option!
> 
> 5.                   “Terrabrick”, allows you to start with one brick
> and expand as you need to.
> 
> 6.         Increased reliability by support for diskless cluster nodes
> 
>  
> 
> Regards
> 
> Imran
> 
> ------------------------------------------------------------------------
> 
> *From:* beowulf-bounces at beowulf.org [mailto:beowulf-bounces at beowulf.org]
> *On Behalf Of *ar 3107
> *Sent:* 10 November 2005 05:36
> *To:* beowulf at beowulf.org
> *Subject:* [Beowulf] Question on hgh performance, low cost Fileserver
> 
>  
> 
> We are looking into designing a low cost, high performance storage
> system. Requirements as below:
> 
> - Starts at 3TB, should scale up by adding more servers to say 10-12TB
> - Use commodity technologies (x86_64, IB, GE, Linux), preferably all OSS
> components
> - Provide high I/O which scales with addition of storage nodes.
> - To be used for hosting user home dirs so reliability is important
> - The HPC cluster starts with 6 AMD64 nodes and is expected to scale to
> 1000+nodes in a year.
> - Preferably without FC/SAN
> 
> We do have experience with IBM GPFS, PVFS (1,2), NetApps, PolyServe but
> not with GFS and LUSTRE.
> 
> PVFS is not reliable enough for home dirs (OK for scratch), GPFS cannot
> do RAID5 like striping across nodes, needs SAN for RAID1 like mirroring
> (cost $$$) , polyserve is too expensive (per CPU pricing)
> 
> Is GFS or Lustre suitable for the above needs? Any other commercial slution?
> 
> I would like to know the experiences and suggestions from the advanced
> users on this list.
> 
> Regards,
> Arvind




More information about the Beowulf mailing list