Hannes Reinecke

> @hreinecke I'm trying to run this test case on my system and observe failure. nvme connect fails and kernel reports an error: > > ``` > [ 3441.943356][ T2668]...

The Megaraid SAS HBAs have a complete SCSI emulation internally, so any command you send will have to be interpreted by firmware. Chances are that it simply doesn't understand SES...

That would depend on the RAID controller. Some controllers (eg aacraid) expose the internal disks to the OS, but do no attach a disk device to them. For those it...

Well, that will be tricky. This is an aacraid controller, which will hide the SCSI drives behind the firmware. Some models have the ability to expose the physical devices (there...

_Actually_ there's not much we can do. Ultimately it's down to the SCSI HBA how many LUNs it can address, and which addressing scheme is supported. Clearly it'll be pointless...

The biggest problem here is that it's anyones guess _how_ LUN 256 will be presented. So it's virtually impossible to predict _what_ the storage array will accept for that LUN....

Well; from the initiator side there is nothing to 'deal' with wrt LUN numbers. We actually need to use them as is as the target might derive some routing information...

Good point. Care to write a json output for sg3_utils?

Note that 'fcloop' currently has some issues, as it'll retry authentication failures, causing repeated connect/disconnect attempts to fail. Patch is pending in the linux-nvme mailing list ('nvme-auth: retry command if...

Patchset submitted to linux-nvme, thread 'nvme: honor DNR status for authentication' to fix the connect behaviour wrt DNR bit handling.