Re: Bigger Libraries (was: Re: Extent SMPPTS Question)
- From: Tom.Schmidt@xxxxxxxxxxxx (Tom Schmidt)
- Date: 15 Dec 2005 21:10:44 -0800
On Thu, 15 Dec 2005 16:42:41 -0700, Paul Gilmartin wrote:
>In a recent note, McKown, John said:
>> Date: Thu, 15 Dec 2005 16:17:56 -0600
>> > -----Original Message-----
>> > From: IBM Mainframe Discussion List
>> > On Behalf Of Paul Gilmartin
>>
>> > IBM's limits on PDS size (also PDSE) are _so_ 20th Century! Individal
>> > program products shouldn't be compelled to such funky circumventions.
>>
>> For the truly weird amoung us, wouldn't it be "interesting" if IBM were
>> to "get around" this problem by allowing SMPPTS to point to an UNIX
>> subdirectory (like SMPNTS does now). This would allow a very large
>> SMPPTS because an HFS or zFS filesystem can span a LOT of volumes. Hum,
>> how big is 59 times a 3390-54 (or whatever the "largest" 3390 volume is
>> now).
>>
>Ah! You expect an IBM product to make the transition to a
>superior technology. Clearly SMP/E saw advantages to UNIX
>filesystems for NTS (not to minimize the troubles reported
>previously in this list when doing enormous RECEIVE FROMNETWORK).
>
>But I always seek the global solution rather than the local.
>This is a problem for DF/SMS, not SMP/E. BPAM support for
>reading UNIX directories already exists; the remaining
>deficiency is write support and that newfangled replacement
>for BLDL (the memory is the second thing to go). Ah! DESERV.
>Completing BPAM support for UNIX would relieve SMP/E of
>most of the coding.
Gil,
Why stop there? With BPAM write/update support for Posix files we could
substitute NFS for HFS or zFS, right? So the SMP/E disk support could be
off-platform (on SAN or NAS?) if the customer so desired. (Where they
might not be backed up reliably... but that's a different topic.)
I don't think that the PDS directory simulation is the only issue that IBM
would have to deal with though. Wouldn't there need to be a conversion
process or does this magically drop in at a new release of z/OS, CICS,
IMS, DB2, etc.? (How many sysprogs or their managers will want to make
that leap of faith???)
I'm not convinced that UNIX file system(s) are superior to z/OS' native
file systems, but I'm all for variety and letting the customer &
marketplace decide. I'd be less convinced if DFSMS allowed for
substantially more than 59 volumes... I used to think that the 255 volume
limit was too small and then IBM lowered it by 77%!
--
Tom Schmidt
Madison, WI
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to listserv@xxxxxxxxxxx with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
.
- Prev by Date: Re: can't grasp the concept of "zero" (was: Why does LMPUT(ISPF Library service)?)
- Next by Date: RE: COND= on an EXEC Statement
- Previous by thread: Bigger Libraries (was: Re: Extent SMPPTS Question)
- Next by thread: RE: Bigger Libraries (was: Re: Extent SMPPTS Question)
- Index(es):
Relevant Pages
|
