Hi,
I am trying to build a SOFS cluster which is backed by two iSCSI NAS devices, but I seem to be having issues if one of the NAS devices fails.
here is what I have (all systems are based on SERVER 2012 R2 Preview):
two NAS devices NAS1 and NAS2, there are two iSCSI disks on each because the SOFS cluster storage pool seems to want to have a minimum of 3 disks, so I’ve got 4 between the 2 NAS.
FS1 and FS2, both file servers have the iSCSI initiator connect to all targets on the NAS - so in each FS I see four attached iSCSI disks in disk management. I create a cluster between FS1 and FS2, create a storage pool in failover cluster manager and include all 4 disks, then I create a virtual disk within this storage pool, finally I add the file server role under the role services section, create a share and assign relevant permissions.
Everything is online, everything looks good. I can transfer the scale out file service role between FS1 and FS2 as I expect, the cluster name for this role is SOFS1.
So to test that the backend disk system can fail, I turn off NAS1 but the storage pool in my cluster shows errors and warnings that disks are not available and I cant bring them online even though both FS1 and FS2 are online. The file services cluster role is showing that its online.
My question here is two part:
Have I completely missed the point of a scale out file server and this is not how its supposed to work (i.e. is it just meant to work if one of the FS goes down rather than the backend NAS)?
Or… is this is unexpected behaviour, is it not just meant to use NAS2?
I assumed that the storage pool I created which included disks from both NAS1 and NAS2 will mirror the data across those disks so removing either NAS shouldn’t fail the storage?
I would appreciate any advice on this matter.
Thanks
Steve