On 03/12/2014 04:23 PM, Dan Mossor issued this missive: > I've been combing the web for a couple days, and most hits I'm finding > regarding using a btrfs subvolume as an iSCSI target are posts from the > 2009/2010 timeframe stating that that functionality is broken. > > I'm working on a couple projects here, and one of those involves the > research into iSCSI targets and initiators, and the testing thereof. I > have an F20 system with a 1.3TB btrfs master subvolume (across three > dissimilar devices), and want to create a subvolume within that to serve > as an iSCSI target. > > The traditional methods, specifically those mentioned in the Fedora Wiki > and the tgtadm man pages, place the emphasis on using devices (/dev/sdx) > or files (dd if=/dev/zero of=/array/target1 bs=1k count=1). I realize > that using the file is probably the easiest thing to do in this > instance, but that's not the point of my project - I am intentionally > trying to designate the btrfs subvolume as the target. > > The issue I believe that is going to stop me is that tgtadm expects a > block device as the target, and btrfs does not provide that. Is there a > workaround, other than using a file in the subvolume as the target? Is > there work in progress to enable this feature, either in the > scsi-target-utils package or within the btrfs framework? iSCSI stands for "internet SCSI", a way to share raw devices over normal network mechanisms rather than having to have a storage farm (e.g. fiberchannel switches or SAS concentrators and a ton of cables). As far as I know, the functionality you want will never be available. iSCSI is intended to share raw devices, not subdirectories (that's what things like NFS and CIFS are for).

I've been combing the web for a couple days, and most hits I'm finding regarding using a btrfs subvolume as an iSCSI target are posts from the 2009/2010 timeframe stating that that functionality is broken. I'm working on a couple projects here, and one of those involves the research into iSCSI targets and initiators, and the testing thereof. I have an F20 system with a 1.3TB btrfs master subvolume (across three dissimilar devices), and want to create a subvolume within that to serve as an iSCSI target. The traditional methods, specifically those mentioned in the Fedora Wiki and the tgtadm man pages, place the emphasis on using devices (/dev/sdx) or files (dd if=/dev/zero of=/array/target1 bs=1k count=1). I realize that using the file is probably the easiest thing to do in this instance, but that's not the point of my project - I am intentionally trying to designate the btrfs subvolume as the target. The issue I believe that is going to stop me is that tgtadm expects a block device as the target, and btrfs does not provide that. Is there a workaround, other than using a file in the subvolume as the target? Is there work in progress to enable this feature, either in the scsi-target-utils package or within the btrfs framework?