Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread David Boyes
> The tape support should be in place to my knowledge. What's missing is
> the tape management piece.

Correct. 

> Of course we could consider writing a driver that applies tape library
> functions for VTS but without any tape management software on top it
> would only be a piece of technology. I.e. if there is any outlook for
> an exploiter we can consider putting in onto the requirements list.

A simple exploiter is 'mtx'. That's where all other Unix/Linux tape
control function lives. Most applications that care about such things
call 'mtx' to do the actual work. 

> But then the tape library API may not be published causing
implementation
> delays and possibly you could instead attach to the VTS through FCP,
too
> and use TSM support off the box. 

Until native SCSI tape support is in the plan for VM, VSE and z/OS, this
will be a hard case to make, even if the device and library can support
multiple interfaces simultaneously. Switching back and forth from FCP to
channel-attached is awkward and operationally a non-starter. 

Of course you could build the interface code into TSM to deal with
channel-attached tape. That's already OCO, and there's your
exploiter...8-)

> That way you wouldn't have to hold your
> breath while we figure out how to deliver on your requirement
validating
> with more customers such that we understand its general market
importance
> and push it into plans eventually - still awaiting tape management
> exploitation ... Thoughts?

See suggestion to base the Linux part of the changer interface on IUCV
to the VM DFSMS or via TCP to z/OS DFSMS. AFAIK, both are published
interfaces, and you can continue to hide the gory details of how to
actually work the device w/o tainting the kernel. 

Possible implementation: provide a back end to the generic SCSI device
(sg) driver to translate the generic SCSI tape management commands into
the right IUCV or TCP primitives for DFSMS. Then you need no user space
mods for existing tape-aware apps, and TSM can just use it as-is. 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Tom Duerbusch

I've done a fair amount of prototyping using different varients of
virtual tape and also trying to use our VSE TLMS and channel connected
tape drives with zLinux.

Basically for zLinux, I send mount requests to a VSE service machine,
that causes a Dynam mount request to be done.  When satisfied, the tape
drive is attached back over to the zLinux machine.

The biggest problem I have, and will also be if we have VTS command
function in zLinux, is how to handle EOV.  And in a VTS, the virtual
volume is a lot smaller than the standard real tape cart.  Perhaps
having the 3590 device driver updated to call an exit at EOV, (perhaps a
rexx exec(?)) to reissue the VTS commands.

The second, smaller problem, is not all software products seem to like
the 3590 tape support for zLinux.
I can "tar" things, but Oracle didn't seem to like it, but it could be
my setup also. 

Right now, I'm on, what the Open Systems people call D2D2T (disk to disk
to tape).

I backup the zLinux image to a disk file.
I FTP the disk file to VSE.
I backup the disk file using VSE utilities and TLMS.

If I had the VSE "Barnard" software IP stack, I could ftp directly to a
tape drive.  We have the CSI stack which doesn't have this function, at
this time.  That would bring it down to D2T.

Also, been experimenting with VSE Virtual Tape with the server back on
zLinux.  Trying to come up with an interm "robotic TLMS" until we get
budgetted for our real VTS.

Life is fun, when you don't have money for a real solution .

I have a drawer full of spit, bubble gum, purple smoke, lasers and duct
tape which is used to put all of this together.

But thinking about have VTS command functions available to zLinux, the
first swipe at it might be interfacing to z/VM to use a CMS server as
the interface to the VTS command functions.  Good way to get something
out there and debug the prototype before moving it to a native zLinux
application.

Well, you wanted comments :)

Tom Duerbusch
THD Consulting

Ingo Adlung wrote:


The tape support should be in place to my knowledge. What's missing is
the tape management piece.

Of course we could consider writing a driver that applies tape library
functions for VTS but without any tape management software on top it
would only be a piece of technology. I.e. if there is any outlook for
an exploiter we can consider putting in onto the requirements list.
But then the tape library API may not be published causing implementation
delays and possibly you could instead attach to the VTS through FCP, too
and use TSM support off the box. That way you wouldn't have to hold your
breath while we figure out how to deliver on your requirement validating
with more customers such that we understand its general market importance
and push it into plans eventually - still awaiting tape management
exploitation ... Thoughts?

Best regards
Ingo





--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Ingo Adlung
The tape support should be in place to my knowledge. What's missing is
the tape management piece.

Of course we could consider writing a driver that applies tape library
functions for VTS but without any tape management software on top it
would only be a piece of technology. I.e. if there is any outlook for
an exploiter we can consider putting in onto the requirements list.
But then the tape library API may not be published causing implementation
delays and possibly you could instead attach to the VTS through FCP, too
and use TSM support off the box. That way you wouldn't have to hold your
breath while we figure out how to deliver on your requirement validating
with more customers such that we understand its general market importance
and push it into plans eventually - still awaiting tape management
exploitation ... Thoughts?

Best regards
Ingo

Linux on 390 Port  wrote on 01.05.2007 20:41:03:

> True.  Except that there also needs to be drivers for the tape and the
> VTS.  And tape management.  I'm guessing that even if all of the
> different IBMs got their act together, it would be awhile before we
> would see something usable.
>
> > -Original Message-
> > From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On
> > Behalf Of Dave Jones
> > Sent: Tuesday, May 01, 2007 1:37 PM
> > To: LINUX-390@VM.MARIST.EDU
> > Subject: Re: Linux on z and VTS == compatible at all?
> >
> > Chris, I think the real roadblock to getting this done lies
> > with Tivoli, and not with IBM itself.
> >
> > DJ
> >
> > Little, Chris wrote:
> > > U.
> > >
> > > IBM?  Are you there?  Can you hear me?  Would like to drive
> > a TS7700
> > > from Linux.  Pretty please?
> > >
> > >> -Original Message-
> > >> From: Linux on 390 Port [mailto:[EMAIL PROTECTED]
> > On Behalf Of
> > >> Dave Jones
> > >> Sent: Tuesday, May 01, 2007 9:25 AM
> > >> To: LINUX-390@VM.MARIST.EDU
> > >> Subject: Re: Linux on z and VTS == compatible at all?
> > >>
> > >> Then, as DB has noted, you're basically SOL.sorry.
> > >>
> > >> DJ
> > >>
> >
> > --
> > For LINUX-390 subscribe / signoff / archive access
> > instructions, send email to [EMAIL PROTECTED] with the
> > message: INFO LINUX-390 or visit
> > http://www.marist.edu/htbin/wlvindex?LINUX-390
> >
>
> --
> For LINUX-390 subscribe / signoff / archive access instructions,
> send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or
visit
> http://www.marist.edu/htbin/wlvindex?LINUX-390

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread David Boyes
> > Would be real nice if we got a device driver for Linux for these
> > beasties... sign. Another requirement in the works.
> I won't say "impossible", but at present, IBM considers the interface
to
> talk to the VTS (and the ATL robot as well) to be proprietary and is
> undocumented. But then, the original 3590 driver was binary-only and
is
> now "open", so maybe some day.

Write it as an AF_IUCV connection to a VM-based daemon, and drive
DFSMS/VM to do the deed. More scalable, and no IP exposure required. Not
even a GPL burp; none of the OCO bits are in Linux at all. 

In fact, if you have a driver for the device type your VTS emulates, see
our mount daemon for VM and Linux. It's an example of how to do this w/o
causing problems.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Little, Chris
I can understand that . . . . 

> -Original Message-
> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On 
> Behalf Of Mark Post
> Sent: Tuesday, May 01, 2007 1:54 PM
> To: LINUX-390@VM.MARIST.EDU
> Subject: Re: Linux on z and VTS == compatible at all?
> 
> >>> On Tue, May 1, 2007 at  2:45 PM, in message
> <[EMAIL PROTECTED]>
> , "Little, Chris"
> <[EMAIL PROTECTED]> wrote: 
> > Wouldn't it be possible to have a public interface and private 
> > non-kernel version/distribution specific code?  I'm certainly not a 
> > kernel hacker, so I p'raps I'm talking out my back-end here.
> 
> It's technically possible, but politically a bad thing.  The 
> kernel developers call them "binary blobs" and there is 
> continuous argument over whether they're legal under the GPL or not.
> 
> 
> Mark Post
> 
> --
> For LINUX-390 subscribe / signoff / archive access 
> instructions, send email to [EMAIL PROTECTED] with the 
> message: INFO LINUX-390 or visit 
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Mark Post
>>> On Tue, May 1, 2007 at  2:45 PM, in message
<[EMAIL PROTECTED]>, "Little, Chris"
<[EMAIL PROTECTED]> wrote: 
> Wouldn't it be possible to have a public interface and private
> non-kernel version/distribution specific code?  I'm certainly not a
> kernel hacker, so I p'raps I'm talking out my back-end here.

It's technically possible, but politically a bad thing.  The kernel developers 
call them "binary blobs" and there is continuous argument over whether they're 
legal under the GPL or not.


Mark Post

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Little, Chris
Wouldn't it be possible to have a public interface and private
non-kernel version/distribution specific code?  I'm certainly not a
kernel hacker, so I p'raps I'm talking out my back-end here.

> -Original Message-
> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On 
> Behalf Of McKown, John
> Sent: Tuesday, May 01, 2007 1:40 PM
> To: LINUX-390@VM.MARIST.EDU
> Subject: Re: Linux on z and VTS == compatible at all?
> 
> > -Original Message-
> > From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On 
> Behalf Of 
> > David Boyes
> > Sent: Tuesday, May 01, 2007 1:29 PM
> > To: LINUX-390@VM.MARIST.EDU
> > Subject: Re: Linux on z and VTS == compatible at all?
> > 
> > 
> > > Channel attached.
> > 
> > Would be real nice if we got a device driver for Linux for these 
> > beasties... sign. Another requirement in the works.
> >  
> 
> I won't say "impossible", but at present, IBM considers the 
> interface to talk to the VTS (and the ATL robot as well) to 
> be proprietary and is undocumented. But then, the original 
> 3590 driver was binary-only and is now "open", so maybe some day.
> 
> 
> > 
> 
> --
> John McKown
> Senior Systems Programmer
> HealthMarkets
> Keeping the Promise of Affordable Coverage Administrative 
> Services Group Information Technology
> 
> The information contained in this e-mail message may be 
> privileged and/or confidential.  It is for intended 
> addressee(s) only.  If you are not the intended recipient, 
> you are hereby notified that any disclosure, reproduction, 
> distribution or other use of this communication is strictly 
> prohibited and could, in certain circumstances, be a criminal 
> offense.  If you have received this e-mail in error, please 
> notify the sender by reply and delete this message without 
> copying or disclosing it. 
> 
> --
> For LINUX-390 subscribe / signoff / archive access 
> instructions, send email to [EMAIL PROTECTED] with the 
> message: INFO LINUX-390 or visit 
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Little, Chris
True.  Except that there also needs to be drivers for the tape and the
VTS.  And tape management.  I'm guessing that even if all of the
different IBMs got their act together, it would be awhile before we
would see something usable.

> -Original Message-
> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On 
> Behalf Of Dave Jones
> Sent: Tuesday, May 01, 2007 1:37 PM
> To: LINUX-390@VM.MARIST.EDU
> Subject: Re: Linux on z and VTS == compatible at all?
> 
> Chris, I think the real roadblock to getting this done lies 
> with Tivoli, and not with IBM itself.
> 
> DJ
> 
> Little, Chris wrote:
> > U.
> >
> > IBM?  Are you there?  Can you hear me?  Would like to drive 
> a TS7700 
> > from Linux.  Pretty please?
> >
> >> -Original Message-
> >> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] 
> On Behalf Of 
> >> Dave Jones
> >> Sent: Tuesday, May 01, 2007 9:25 AM
> >> To: LINUX-390@VM.MARIST.EDU
> >> Subject: Re: Linux on z and VTS == compatible at all?
> >>
> >> Then, as DB has noted, you're basically SOL.sorry.
> >>
> >> DJ
> >>
> 
> --
> For LINUX-390 subscribe / signoff / archive access 
> instructions, send email to [EMAIL PROTECTED] with the 
> message: INFO LINUX-390 or visit 
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread McKown, John
> -Original Message-
> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On 
> Behalf Of David Boyes
> Sent: Tuesday, May 01, 2007 1:29 PM
> To: LINUX-390@VM.MARIST.EDU
> Subject: Re: Linux on z and VTS == compatible at all?
> 
> 
> > Channel attached.
> 
> Would be real nice if we got a device driver for Linux for these
> beasties... sign. Another requirement in the works. 
>  

I won't say "impossible", but at present, IBM considers the interface to
talk to the VTS (and the ATL robot as well) to be proprietary and is
undocumented. But then, the original 3590 driver was binary-only and is
now "open", so maybe some day.


> 

--
John McKown
Senior Systems Programmer
HealthMarkets
Keeping the Promise of Affordable Coverage
Administrative Services Group
Information Technology

The information contained in this e-mail message may be privileged
and/or confidential.  It is for intended addressee(s) only.  If you are
not the intended recipient, you are hereby notified that any disclosure,
reproduction, distribution or other use of this communication is
strictly prohibited and could, in certain circumstances, be a criminal
offense.  If you have received this e-mail in error, please notify the
sender by reply and delete this message without copying or disclosing
it. 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Dave Jones

Chris, I think the real roadblock to getting this done lies with Tivoli,
and not with IBM itself.

DJ

Little, Chris wrote:

U.

IBM?  Are you there?  Can you hear me?  Would like to drive a TS7700
from Linux.  Pretty please?


-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On
Behalf Of Dave Jones
Sent: Tuesday, May 01, 2007 9:25 AM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: Linux on z and VTS == compatible at all?

Then, as DB has noted, you're basically SOL.sorry.

DJ



--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Little, Chris
U.

IBM?  Are you there?  Can you hear me?  Would like to drive a TS7700
from Linux.  Pretty please? 

> -Original Message-
> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On 
> Behalf Of Dave Jones
> Sent: Tuesday, May 01, 2007 9:25 AM
> To: LINUX-390@VM.MARIST.EDU
> Subject: Re: Linux on z and VTS == compatible at all?
> 
> Then, as DB has noted, you're basically SOL.sorry.
> 
> DJ
> 
> Little, Chris wrote:
> > Channel attached.
> >
> > We're running TSM on z/OS.  With the level of utilization, though, 
> > it's getting to be expensive.  We were looking to cut costs 
> by moving 
> > the server to engines with flat pricing.
> >
> >> -Original Message-
> >> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] 
> On Behalf Of 
> >> David Boyes
> >> Sent: Monday, April 30, 2007 7:46 PM
> >> To: LINUX-390@VM.MARIST.EDU
> >> Subject: Re: Linux on z and VTS == compatible at all?
> >>
> >>> What is the possibility of reliably running TSM server on 
> Linux with
> >> the
> >>> backups going to our brand spanking new clustered TS7700?
> >>> I couldn't find any literature supporting it.
> >> If it's channel-attached, then the probability is exactly 
> zero. TSM 
> >> on Linux doesn't do channel-attached tape. If you had TSM 
> on z/OS -- 
> >> or even the old TSM on VM server -- you could drive it 
> that way using 
> >> DFSMS/VM RMM and use server-to-server migration to get the TSM 
> >> servers to work together.
> >>
> >> I don't know of anyone using it via FCP attached tape. The generic 
> >> SCSI changer commands don't seem to work very reliably; 
> the only SCSI 
> >> VTS I have access to seems to hang reliably when presented 
> with SCSI 
> >> changer commands. This makes the PTBs unhappy, so I've stopped 
> >> trying.
> >>
> >> 
> -
> >> - For LINUX-390 subscribe / signoff / archive access instructions, 
> >> send email to [EMAIL PROTECTED] with the
> >> message: INFO LINUX-390 or visit
> >> http://www.marist.edu/htbin/wlvindex?LINUX-390
> >>
> >
> > 
> --
> > For LINUX-390 subscribe / signoff / archive access 
> instructions, send 
> > email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or 
> > visit http://www.marist.edu/htbin/wlvindex?LINUX-390
> 
> --
> DJ
> V/Soft
> 
> --
> For LINUX-390 subscribe / signoff / archive access 
> instructions, send email to [EMAIL PROTECTED] with the 
> message: INFO LINUX-390 or visit 
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread David Boyes
> Channel attached.

Would be real nice if we got a device driver for Linux for these
beasties... sign. Another requirement in the works. 
 
> We're running TSM on z/OS.  With the level of utilization, though,
it's
> getting to be expensive.  We were looking to cut costs by moving the
> server to engines with flat pricing.

If you can afford both versions, then move the client interaction to the
TSM on Linux server (ie, let that be the part that talks to the
individual clients) and do only level 2->N migration from the Linux TSM
to the z/OS TSM. It's still not perfect, but you should see your CPU
utilization on z/OS TSM drop substantially. 

You might also compare the price of TSM/VM vs TSM/zOS. If all you use it
for is to drive the tapes, it might be cheaper to license it on VM, and
provide access to the drives via DFSMS/VM talking to DFSMS on z/OS to
get drives allocated and loaded. 

The big catch to this approach is that TSM/VM is supported, but barely
so, and they'd like it to go away entirely. 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Dave Jones

Then, as DB has noted, you're basically SOL.sorry.

DJ

Little, Chris wrote:

Channel attached.

We're running TSM on z/OS.  With the level of utilization, though, it's
getting to be expensive.  We were looking to cut costs by moving the
server to engines with flat pricing.


-Original Message-
From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On
Behalf Of David Boyes
Sent: Monday, April 30, 2007 7:46 PM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: Linux on z and VTS == compatible at all?


What is the possibility of reliably running TSM server on Linux with

the

backups going to our brand spanking new clustered TS7700?
I couldn't find any literature supporting it.

If it's channel-attached, then the probability is exactly
zero. TSM on Linux doesn't do channel-attached tape. If you
had TSM on z/OS -- or even the old TSM on VM server -- you
could drive it that way using DFSMS/VM RMM and use
server-to-server migration to get the TSM servers to work together.

I don't know of anyone using it via FCP attached tape. The
generic SCSI changer commands don't seem to work very
reliably; the only SCSI VTS I have access to seems to hang
reliably when presented with SCSI changer commands. This
makes the PTBs unhappy, so I've stopped trying.

--
For LINUX-390 subscribe / signoff / archive access
instructions, send email to [EMAIL PROTECTED] with the
message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390



--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


--
DJ
V/Soft

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-05-01 Thread Little, Chris
Channel attached.

We're running TSM on z/OS.  With the level of utilization, though, it's
getting to be expensive.  We were looking to cut costs by moving the
server to engines with flat pricing.

> -Original Message-
> From: Linux on 390 Port [mailto:[EMAIL PROTECTED] On 
> Behalf Of David Boyes
> Sent: Monday, April 30, 2007 7:46 PM
> To: LINUX-390@VM.MARIST.EDU
> Subject: Re: Linux on z and VTS == compatible at all?
> 
> > What is the possibility of reliably running TSM server on Linux with
> the
> > backups going to our brand spanking new clustered TS7700?
> > I couldn't find any literature supporting it.
> 
> If it's channel-attached, then the probability is exactly 
> zero. TSM on Linux doesn't do channel-attached tape. If you 
> had TSM on z/OS -- or even the old TSM on VM server -- you 
> could drive it that way using DFSMS/VM RMM and use 
> server-to-server migration to get the TSM servers to work together.
> 
> I don't know of anyone using it via FCP attached tape. The 
> generic SCSI changer commands don't seem to work very 
> reliably; the only SCSI VTS I have access to seems to hang 
> reliably when presented with SCSI changer commands. This 
> makes the PTBs unhappy, so I've stopped trying.
> 
> --
> For LINUX-390 subscribe / signoff / archive access 
> instructions, send email to [EMAIL PROTECTED] with the 
> message: INFO LINUX-390 or visit 
> http://www.marist.edu/htbin/wlvindex?LINUX-390
> 

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-04-30 Thread David Boyes
> What is the possibility of reliably running TSM server on Linux with
the
> backups going to our brand spanking new clustered TS7700?
> I couldn't find any literature supporting it.

If it's channel-attached, then the probability is exactly zero. TSM on
Linux doesn't do channel-attached tape. If you had TSM on z/OS -- or
even the old TSM on VM server -- you could drive it that way using
DFSMS/VM RMM and use server-to-server migration to get the TSM servers
to work together.

I don't know of anyone using it via FCP attached tape. The generic SCSI
changer commands don't seem to work very reliably; the only SCSI VTS I
have access to seems to hang reliably when presented with SCSI changer
commands. This makes the PTBs unhappy, so I've stopped trying.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Re: Linux on z and VTS == compatible at all?

2007-04-30 Thread Tom Duerbusch
The keyword is "reliably".  Noop.

But if you find something that really works with it, I'm all ears.

So far, the VTS seems to be a 390 thingie.  No FCP support.
Linux applications seem to like scsi tape drives (FCP) not FICON/ESCON.

Then there is the point of control.  That is being able to send mount info to 
the VTS.  390 side so far.

Yep, you can use the VTS on the 390 side to mount tapes and then give the tape 
drive to Linuxand it sort of works.

Biggest problem is handling EOV conditions.  That is way more code then I 
really wanted to handle.
The virtual tapes on a VTS, are a lot smaller then a physical tape, which 
increases the problem of EOV conditions.

Yep, much of this can be handled by more and more local code.  And the more you 
code, the less and less it will be tested.  So, reliability decreases.

Right now, I toying with the idea of taking Linux backups to disk, and then ftp 
the backups to the 390 side to be staged to tape.  The open system side calls 
this D2D2T (disk to disk to tape).  Not sure about that, but that is what 
testing is for .

We have a VTS planned for (next budget yr), so I'm testing the process without 
the VTS.

Tom Duerbusch
THD Consulting



>>> "Little, Chris" <[EMAIL PROTECTED]> 4/30/2007 5:28 PM >>>
What is the possibility of reliably running TSM server on Linux with the
backups going to our brand spanking new clustered TS7700?

I couldn't find any literature supporting it.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390


Linux on z and VTS == compatible at all?

2007-04-30 Thread Little, Chris
What is the possibility of reliably running TSM server on Linux with the
backups going to our brand spanking new clustered TS7700?

I couldn't find any literature supporting it.

--
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390