Definitely. The SCSI code predates the bdev UUIDs and was never updated. Do you want to take a crack at some patches to fix that up? If you're not able, we can put something on the Trello backlog.

Thanks, -Jim On 5/31/19, 1:50 PM, "SPDK on behalf of Andrey Kuzmin" < spdk-bounces(a)lists.01.org on behalf of akuzmin(a)jetstreamsoft.com&gt; wrote: It appears that scsi bdev uses bdev name as an input to construct the NAA device identifier under device identification VPD ( https://github.com/spdk/spdk/blob/master/lib/scsi/scsi_bdev.c#L289). Won't it be more reliable to utilize bdev UUID if available? Thanks, Andrey _______________________________________________ SPDK mailing list SPDK(a)lists.01.org https://lists.01.org/mailman/listinfo/spdk _______________________________________________ SPDK mailing list SPDK(a)lists.01.org https://lists.01.org/mailman/listinfo/spdk

Please note that we rely on the bdev name as the device identifier. We expose our host bdevs in guests via virtio-scsi which creates entries under /dev/disk/by-* using the device identifier for the persistent block device naming in the guest.

If we do make a change, please preserve backwards compatibility. In addition to using the UUID, it may be best to also have an RPC option where we can specify the id externally.

— Michael > On May 31, 2019, at 3:37 PM, Andrey Kuzmin <akuzmin(a)jetstreamsoft.com&gt; wrote: > > On Fri, May 31, 2019, 23:54 Harris, James R <james.r.harris(a)intel.com&gt; > wrote: > >> Definitely. The SCSI code predates the bdev UUIDs and was never updated. >> Do you want to take a crack at some patches to fix that up? If you're not >> able, we can put something on the Trello backlog. >> > > Put it on the backlog so that I or anyone else can take on that as time > permits. > > Thanks, > Andrey > >> >> Thanks, >> >> -Jim >> >> >> On 5/31/19, 1:50 PM, "SPDK on behalf of Andrey Kuzmin" < >> spdk-bounces(a)lists.01.org on behalf of akuzmin(a)jetstreamsoft.com&gt; wrote: >> >> It appears that scsi bdev uses bdev name as an input to construct the >> NAA >> device identifier under device identification VPD ( >> https://github.com/spdk/spdk/blob/master/lib/scsi/scsi_bdev.c#L289). >> Won't >> it be more reliable to utilize bdev UUID if available? >> >> Thanks, >> Andrey >> _______________________________________________ >> SPDK mailing list >> SPDK(a)lists.01.org >> https://lists.01.org/mailman/listinfo/spdk >> >> >> _______________________________________________ >> SPDK mailing list >> SPDK(a)lists.01.org >> https://lists.01.org/mailman/listinfo/spdk >> > _______________________________________________ > SPDK mailing list > SPDK(a)lists.01.org > https://lists.01.org/mailman/listinfo/spdk _______________________________________________ SPDK mailing list SPDK(a)lists.01.org https://lists.01.org/mailman/listinfo/spdk