[Stgt-devel] tape drive not supported yet?

Mark Harvey markh794
Thu Jul 17 07:07:02 CEST 2008


On Thu, Jul 17, 2008 at 2:57 PM, FUJITA Tomonori
<fujita.tomonori at lab.ntt.co.jp> wrote:
> On Wed, 16 Jul 2008 16:51:38 +0900
> FUJITA Tomonori <fujita.tomonori at lab.ntt.co.jp> wrote:
>
>> On Sat, 12 Jul 2008 19:40:21 +0200
>> Albert Pauw <albert.pauw at gmail.com> wrote:
>>
>> > Tried to set up a tape drive target:
>> >
>> > LUN=$(($LUN+1))
>> > ID=Tape
>> > SN=54321
>> > tgtadm --lld iscsi --op new --mode logicalunit --tid $TID --lun $LUN \
>> > --device-type tape
>> >
>> > Unfortunately, tgtadm tells me:
>> >
>> > tgtadm: type emulation isn't supported yet
>> >
>> > Is this correct?
>>
>> You are talking about vtl, right?
>>
>> Here's aftab azmi's vtl patch:
>>
>> https://lists.berlios.de/pipermail/stgt-devel/2008-May/001610.html
>>
>> I've cleaned up it and my initiator can find a tape drive with this
>> pathc. However, seems that this patch has several issues to solve.
>>
>> Mark's comments are:
>>
>> https://lists.berlios.de/pipermail/stgt-devel/2008-May/001622.html
>>
>> I need vtl to just work on the tape initiator driver. I'll try to
>> improve the vtl support but probabaly I don't have time to make the
>> vtl support good enough for real usage.
>
> BTW, I'll merge this patch unless someone is against it.

No arguments from my end.

I've been attempting to get what I have in some sort of state that will:
a) compile cleanly
b) have some basic structure to work from.
c) have something in basic (read block / write block of data) working order.

Just when I find time to get point a), I rebase and am back to
cleaning up the conflicts..
Then I look over what I've done and decide that the code should not be
inflicted on anybody :)

>
> As discussed, it's not finished at all, but it's easier to work with
> the in-tree code rather than out-of-tree code.
>

Agreed.

Cheers
Mark



More information about the stgt mailing list