Aaron,
yes its possible - couple of thins you need to consider
JBOD means that there is no data resiliency at a hardware level
this means that you need to have some level of software based data resiliency
Typically this implies either 2 or more copies of the data (on independent resilient hardware)
in your scenario this would typically mean 1/2 the drives from JBOD1 + 1/2 the drives from JBOD2
configured into Filesystem 1
1/2 the drives from JBOD 1 + 1/2 the Drives from JBOD2 configured into Filesystem2
the net result being that you will end up with 50% of your available capacity presented to your Filesystems
if you want more information on how this works look up FPO in the Spectrum Scale Knowledge Centre
The other scenario that you can consider is if you look at Scale Erasure Code Edition (Scale ECE)
you need a minimum of 4 Servers for Scale ECE but with Scale ECE you do get Raid 4+2P, Raid 4+3P
Raid 8+2P and Raid8+3P options, as well as 2 copy and 3copy options -- you would need to reach out to your local IBM team to validate if Scale ECE would be supported with JBOD shelves presented to multiple servers.
Regards
Andrew Beattie
File and Object Storage Technical Specialist - A/NZ
IBM Systems - Storage
Phone: 614-2133-7927
E-mail: abeat...@au1.ibm.com
----- Original message -----
From: Aaron Turner <a.tur...@lboro.ac.uk>
Sent by: gpfsug-discuss-boun...@spectrumscale.org
To: "gpfsug-discuss@spectrumscale.org" <gpfsug-discuss@spectrumscale.org>
Cc:
Subject: [EXTERNAL] [gpfsug-discuss] Identifiable groups of disks?
Date: Tue, May 14, 2019 6:47 PM
Scenario:
- one set of JBODS
- want to create two GPFS file systems
- want to ensure that file system A uses physical disks a0, a1... an-1 and file system B uses physical disks b0, b1... bn-1
- want to be able to assign specific sets of disks a0..an-1, b0..bn-1 on creation
- Potentially allows all disks b0..bn-1 to be destroyed if required whilst not affecting a0..an-1
Is this possible in GPFS?Regards______________________________________Aaron TurnerSenior IT Services Specialist in High Performance ComputingLoughborough Universitya.tur...@lboro.ac.uk01509 226185_____________________________________________________________________________________
gpfsug-discuss mailing list
gpfsug-discuss at spectrumscale.org
http://gpfsug.org/mailman/listinfo/gpfsug-discuss
_______________________________________________ gpfsug-discuss mailing list gpfsug-discuss at spectrumscale.org http://gpfsug.org/mailman/listinfo/gpfsug-discuss
