Ang: Re: Ang: Re: [Stgt-devel] Re: [Iscsitarget-devel] stgt a new version of iscsi target?
johan at capvert.se
johan at capvert.se
Wed Dec 7 15:28:17 CET 2005
This is a multipart message in MIME format.
--=_alternative 005147A2C12570D0_=
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Hi! =20
Well, I=B4m happy with QLA and LSI. That would be great! But I=B4m sure I =
could get the LSI 10Gbe iScsi HBA target F/W and sdk, when they release=20
the board.
Any comments from Mike?
Best regards from/Med v=E4nliga h=E4lsningar fr=E5n
Johan Kragsterman
http://www.capvert.se
Ming Zhang <mingz at ele.uri.edu>
2005-12-07 15:11
S=E4nd svar till mingz
=20
Till: johan at capvert.se
Kopia: Mike Christie <michaelc at cs.wisc.edu>,=20
iscsitarget-devel at lists.sourceforge.net, stgt=20
<stgt-devel at lists.berlios.de>, stgt-devel-admin at berlios.de, Robert=20
Whitehead <WRWHITEHEAD at novell.com>
=C4rende: Re: Ang: Re: [Stgt-devel] Re: [Iscsitarget-devel] stgt a =
new version of=20
iscsi target?
my 2 cents
most iscsi hba are still closed sourced, at least they only provide
source code sdk to customers or oem developers. so it will be hard to
integrate them unless their own engr do it.
toe is not here, not in scsi domain. toe will be treated as NIC and use
software iscsi target stack coming with stgt. nic is same thing here.
fc/sas, or maybe old parallel scsi will integrate with stgt just like
what qla + scst does now.=20
i remember Mike mentioned that he is working on a qla+stgt. also lsi
+stgt is claimed.
ming
On Wed, 2005-12-07 at 13:51 +0100, johan at capvert.se wrote:
> Hi, Mike!
>=20
>=20
> I, and I suppose a lot of other people, would like know how you look at=20
the
> HW support on the Fc side? I don=B4t know what you mean now, when you talk
> about HW target support, if you talk iScsi HBA, TOE HBA, NIC, or Fc HBA, =
or
> all of them. Since I=B4m very much on the Fc side, I=B4m especially=20
interested
> in your developement strategy for Fc. How is your priority list look=20
like?
>=20
> Best regards from/Med v=E4nliga h=E4lsningar fr=E5n
>=20
> Johan Kragsterman
>=20
> http://www.capvert.se
>=20
>=20
> =20
> Mike Christie =20
> <michaelc at cs.wisc Till: mingz at ele.uri.edu =
=20
> .edu> Kopia: Robert Whitehead=20
<WRWHITEHEAD at novell.com>,=20
> S=E4nt av: iscsitarget-devel at lists.sourceforge.net,=20
stgt=20
> stgt-devel-admin@ <stgt-devel at lists.berlios.de> =
=20
> berlios.de =C4rende: Re: [Stgt-devel] Re=
:=20
[Iscsitarget-devel] stgt a new version of=20
> iscsi target? =20
> =20
> 2005-12-07 11:00 =20
> =20
> =20
>=20
>=20
>=20
>=20
> Ming Zhang wrote:
> > On Tue, 2005-12-06 at 10:58 -0700, Robert Whitehead wrote:
> >
> >>>>>>Ming Zhang <mingz at ele.uri.edu> 12/6/2005 10:32 am >>>
> >>>
> >>>on Tue, 2005-12-06 at 09:56 -0700, Robert Whitehead wrote:
> >>>
> >>>>In a couple of the iscsitarget mail threads there is the reference
> >>
> >>to
> >>
> >>>>the new stgt implementation which is hosted on the
> >>>>http://developer.berlios.de web site. In addition, it also looks
> >>>>like the lasted source code for the IET project has been moved to
> >>
> >>an
> >>
> >>>>iscsitarget project hosted on the developer.berlios.de site. In
> >>
> >>fact
> >>
> >>>>doing a cvs browse on this website (sourceforge) doesn't work and
> >>
> >>in
> >>
> >>>>order to see the sourcecode and you have to go to the berlios
> >>
> >>website to
> >>
> >>>>get the latest iscsitarget code.
> >>>>
> >>>>My question is what is the future of the IET project?
> >>>>
> >>>
> >>>feature freeze and keep it stable.
> >>>
> >>>
> >>>
> >>>
> >>>>It appears the stgt is the new direction. If so what is the
> >>
> >>motivation
> >>
> >>>>for stgt?
> >>>>
> >>
> >>>stgt is a middle layer for different transports. so u can have iscsi,
> >>>fc, sas,... later under same umbrella.
> >>
> >>>ming
> >>
> >>What is the time frame for having initial release of stgt?
> >>Is it is a good state that if I download stgt and build it, will it
> >>function as an iscsi target now?
> >>
> >>Robert
> >>
> >
> >
> > i am not a stgt developer, so i could not tell u the time frame. i=20
think
> > it still need some work before it becomes as stable as iet.
> >
>=20
> Yeah it is very unstable at the moment. We are in the middle if adding
> HW target support so probably scst or IET alone if you just need iscsi
> is better. We have also been trying to figure out a way to reuse some of
> the block layer code for things like barriers/MSG=5F*, tags, requests and
> queue limits etc and that is really a experiment at the moment.
> =5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F
> Stgt-devel mailing list
> Stgt-devel at lists.berlios.de
> http://lists.berlios.de/mailman/listinfo/stgt-devel
>=20
>=20
>=20
>=20
--=_alternative 005147A2C12570D0_=
Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
<br><font size=3D2 face=3D"sans-serif"> &=
nbsp; =
Hi! &=
nbsp; =
<br>
<br>
Well, I=B4m happy with QLA and LSI. That would be great! But I=B4m sure I =
could get the LSI 10Gbe iScsi HBA target F/W and sdk, when they release the=
board.</font>
<br>
<br><font size=3D2 face=3D"sans-serif">Any comments from Mike?</font>
<br><font size=3D2 face=3D"sans-serif"><br>
Best regards from/Med v=
=E4nliga h=E4lsningar fr=E5n<br>
<br>
&nbs=
p; Johan Kragsterman<br>
<br>
&nbs=
p; http://www.capvert.se</font>
<br>
<br>
<br>
<table width=3D100%>
<tr valign=3Dtop>
<td>
<td><font size=3D1 face=3D"sans-serif"><b>Ming Zhang <mingz at ele.uri.edu&=
gt;</b></font>
<p><font size=3D1 face=3D"sans-serif">2005-12-07 15:11</font>
<br><font size=3D1 face=3D"sans-serif">S=E4nd svar till mingz</font>
<br>
<td><font size=3D1 face=3D"Arial"> </font>
<br><font size=3D1 face=3D"sans-serif"> Till: &n=
bsp; johan at capvert.se</font>
<br><font size=3D1 face=3D"sans-serif"> Kopia: &=
nbsp; Mike Christie <michaelc at cs.wisc.edu>, iscsi=
target-devel at lists.sourceforge.net, stgt <stgt-devel at lists.berlios.de>=
;, stgt-devel-admin at berlios.de, Robert Whitehead <WRWHITEHEAD at novell.com=
></font>
<br><font size=3D1 face=3D"sans-serif"> =C4rende=
: Re: Ang: Re: [Stgt-devel] Re: [Iscsitarget-dev=
el] stgt a new version of iscsi target?</font></=
table>
<br>
<br>
<br><font size=3D2 face=3D"Courier New">my 2 cents<br>
<br>
most iscsi hba are still closed sourced, at least they only provide<br>
source code sdk to customers or oem developers. so it will be hard to<br>
integrate them unless their own engr do it.<br>
<br>
toe is not here, not in scsi domain. toe will be treated as NIC and use<br>
software iscsi target stack coming with stgt. nic is same thing here.<br>
<br>
fc/sas, or maybe old parallel scsi will integrate with stgt just like<br>
what qla + scst does now. <br>
<br>
i remember Mike mentioned that he is working on a qla+stgt. also lsi<br>
+stgt is claimed.<br>
<br>
<br>
ming<br>
<br>
<br>
<br>
On Wed, 2005-12-07 at 13:51 +0100, johan at capvert.se wrote:<br>
> =
Hi, M=
ike!<br>
> <br>
> <br>
> I, and I suppose a lot of other people, would like know how you look a=
t the<br>
> HW support on the Fc side? I don=B4t know what you mean now, when you =
talk<br>
> about HW target support, if you talk iScsi HBA, TOE HBA, NIC, or Fc HB=
A, or<br>
> all of them. Since I=B4m very much on the Fc side, I=B4m especially in=
terested<br>
> in your developement strategy for Fc. How is your priority list look l=
ike?<br>
> <br>
> Best regards fr=
om/Med v=E4nliga h=E4lsningar fr=E5n<br>
> <br>
> =
Johan Kragsterman<br>
> <br>
> =
http://www.capvert.se<br>
> <br>
> <br>
> =
 =
; &nb=
sp; &=
nbsp; =
<br>
> =
Mike Christie  =
; &nb=
sp; &=
nbsp; =
<br>
> =
<michaelc at cs.wisc Till: mingz at ele.uri.edu &n=
bsp; =
 =
; <br>
> =
.edu> Kop=
ia: Robert Whitehead <WRWHITEHEAD at novell.com>, &n=
bsp; <br>
> =
S=E4nt av: iscsitar=
get-devel at lists.sourceforge.net, stgt &n=
bsp; <br>
> =
stgt-devel-admin@ <stgt-devel at lists.berlios.d=
e> =
&nbs=
p; <br>
> =
berlios.de =C4rende: Re: [S=
tgt-devel] Re: [Iscsitarget-devel] stgt a new version of <br>
> =
 =
; iscsi target? &nbs=
p; &n=
bsp; =
<br>
> =
 =
; &nb=
sp; &=
nbsp; =
<br>
> =
2005-12-07 11:00 &n=
bsp; =
 =
; &nb=
sp;<br>
> =
 =
; &nb=
sp; &=
nbsp; =
<br>
> =
 =
; &nb=
sp; &=
nbsp; =
<br>
> <br>
> <br>
> <br>
> <br>
> Ming Zhang wrote:<br>
> > On Tue, 2005-12-06 at 10:58 -0700, Robert Whitehead wrote:<br>
> ><br>
> >>>>>>Ming Zhang <mingz at ele.uri.edu> 12/6/2005=
10:32 am >>><br>
> >>><br>
> >>>on Tue, 2005-12-06 at 09:56 -0700, Robert Whitehead wrote:=
<br>
> >>><br>
> >>>>In a couple of the iscsitarget mail threads there is t=
he reference<br>
> >><br>
> >>to<br>
> >><br>
> >>>>the new stgt implementation which is hosted on the<br>
> >>>>http://developer.berlios.de web site. In additio=
n, it also looks<br>
> >>>>like the lasted source code for the IET project has be=
en moved to<br>
> >><br>
> >>an<br>
> >><br>
> >>>>iscsitarget project hosted on the developer.berlios.de=
site. In<br>
> >><br>
> >>fact<br>
> >><br>
> >>>>doing a cvs browse on this website (sourceforge) doesn=
't work and<br>
> >><br>
> >>in<br>
> >><br>
> >>>>order to see the sourcecode and you have to go to the =
berlios<br>
> >><br>
> >>website to<br>
> >><br>
> >>>>get the latest iscsitarget code.<br>
> >>>><br>
> >>>>My question is what is the future of the IET project?<=
br>
> >>>><br>
> >>><br>
> >>>feature freeze and keep it stable.<br>
> >>><br>
> >>><br>
> >>><br>
> >>><br>
> >>>>It appears the stgt is the new direction. If so =
what is the<br>
> >><br>
> >>motivation<br>
> >><br>
> >>>>for stgt?<br>
> >>>><br>
> >><br>
> >>>stgt is a middle layer for different transports. so u can =
have iscsi,<br>
> >>>fc, sas,... later under same umbrella.<br>
> >><br>
> >>>ming<br>
> >><br>
> >>What is the time frame for having initial release of stgt?<br>
> >>Is it is a good state that if I download stgt and build it, wi=
ll it<br>
> >>function as an iscsi target now?<br>
> >><br>
> >>Robert<br>
> >><br>
> ><br>
> ><br>
> > i am not a stgt developer, so i could not tell u the time frame. =
i think<br>
> > it still need some work before it becomes as stable as iet.<br>
> ><br>
> <br>
> Yeah it is very unstable at the moment. We are in the middle if adding=
<br>
> HW target support so probably scst or IET alone if you just need iscsi=
<br>
> is better. We have also been trying to figure out a way to reuse some =
of<br>
> the block layer code for things like barriers/MSG=5F*, tags, requests =
and<br>
> queue limits etc and that is really a experiment at the moment.<br>
> =5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=
=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F=5F<br>
> Stgt-devel mailing list<br>
> Stgt-devel at lists.berlios.de<br>
> http://lists.berlios.de/mailman/listinfo/stgt-devel<br>
> <br>
> <br>
> <br>
> <br>
<br>
</font>
<br>
<br>
--=_alternative 005147A2C12570D0_=--
-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems? Stop! Download the new AJAX search engine that makes
searching your log files as easy as surfing the web. DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
_______________________________________________
Iscsitarget-devel mailing list
Iscsitarget-devel at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
More information about the stgt
mailing list