Adam Thornton wrote:
On May 19, 2008, at 12:02 PM, Alan Ackerman wrote:
One of my managers told me that since you could make both ECKD (FICON)
and SCSI (FCP) connections to the same IBM Storage subsystem, z/Linux
should be able to read z/OS data off the z/OS volumes, without any
special formatting by z/OS. I asked IBM and they said it couldn't be
done - z/Linux cannot read z/OS data and vice-versa.
Is this correct?
If so, what would it take to make z/Linux able to read z/OS data
directly? New drivers? A new file system? How hard would this be to
write?
A new file system device driver, or plug-in for FUSE. I don't know
how hard it would be. I remember someone a few years ago had written
such a thing to read, uh, some kind of MVS files. (I can't even spell
MVS, so don't look to me for help) ISAM? VSAM? Don't remember. I
think he was trying to find a buyer, don't know if he ever succeeded,
and I don't remember his name, but it wasn't one of the regulars on
the Linux/390 list.
I suspect it's pretty hard.
Or hard to sell.
It's over 20 years since I looked closely at this stuff, and it's moved
since then (often accompanying the addition of an "E" to the relevant
acronym).
I would think a filesystem driver would need excp-level access to an
"unformatted" volume. I expect the existing DASD driver gives that, and
that's what the partitioning tools use.
It would need to read HDR1 (I think, but close enough) to discover the
VTOC's location, and read the format4 DSCB. That describes the old
(OS/360-style) datasets.
To access those, you need equivalents to IBM's BSAM/BDAM to read
sequential and direct-access datasets, ISAM, "I don't remember a name"
for partitioned datasets.
There's a fair bit of work there, but in principle it's not hard and
it's documented, at least to the point everything got Enhanced.
VSAM is another matter, all the VTOC describes is data spaces. Some
sites ruled "Each VSAM dataset has its own dataspace," and that helps,
but AFAICR you still need the relevant VSAM catalog and that might be on
another volume.
And I've not addressed whatever magic recent MVS does to cope with those
who'd have datasets extending across volumes: OS/360 and early OS/VS had
multivolume datasets on DASD, but I would be entirely unsurprised to
learn there are different, better ways to achieve that these days.
--
Cheers
John
-- spambait
[EMAIL PROTECTED] [EMAIL PROTECTED]
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)
----------------------------------------------------------------------
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
