Hi,

Have you installed GPFS already? I think the problem might be with the OS rather than with ScaleIO - when you don't use clustering software/filesystem, you are not supposed to share any devices, as the changes done on one node might not be reflected on another.

Can you please deploy the distributed file system and let us know if that worked?

Thank you,

Pawel

GPFS is installed.

As I said if we use iSCSI devices it works fine, same if we add NetApp devices via FC.

This only fails when using SIO as the device.

We have Windows clusters that work correctly.  We assign a quorum disk and 2 data disks, create the cluster and it does what we expect it to do..  formats the devices on one node and all nodes reflect the changes.

There is nothing that was done to the Windows installs other than the drv_cfg command, which we are also using on the RHEL systems.

Hi,

Can you please show us the GPFS status? "mmgetstate -a" and "mmlsmount all -L"

Thanks,

Pawel

The LS Mount status won't matter as the devices fail initialization of GPFS because the systems are not sync'ing the devices across nodes.

I currently have multiple iSCSI devices mounted and mapping correctly.

It's only ScaleIO devices that do not do so.

I think we have some customers using GPFS with ScaleIO without any issues, I am not aware of any special procedure of installation - can you please provide some more details about what are you exactly doing (format etc.) and when exactly GPFS installation fails - what's the command and the error message?

There is no error message, the GPFS cluster simply reports the drive is not available for use.

For now we're simply using iSCSI because we have failed to get ScaleIO to work, at all.

Even after upgrading to Version: 2.0.6035