Terry,

In case you're coming from the z/OS side of the house, each CMS minidisk 
contains a File Status Table (FST), pointing to every file on the disk 
(think of it as the equivalent to the z/OS VTOC).  Each time a minidisk is 
ACCESSed, the FST is read into that user's CMS memory for high-speed 
access.  When a user with R/W access to the disk changes a file (or issues 
"RELEASE"), the in-memory FST is committed to disk, as well. 

But if other userid LINKs the disk MW (multi-write), and then issues the 
ACCESS command for that mdisks, then that other user will read the active 
FST from the disk into their own CMS storage.  When either userid makes 
any change to a file on the disk, their in-storage FST is re-written to 
the disk, but the other user is unaware of the change and their in-memory 
FST is out-of-date.  I call that "one-way encryption".   Things get very 
messy, very quickly.

Precisely what was changed, and the amount of changes will determine how 
badly the minidisk is corrupted and how many complete files may be able to 
be copied from the damaged minidisk to a different minidisk. 

CMS users can and do reliably share a minidisk read-only.  Changes can be 
made to the disk by one R/W user - but all the R/O users then need to 
ACCESS the disk again to "see" any changed (when the ACCESS command 
reloads the FST from the disk into their CMS memory). 

There are manual procedures that permit such shared R/O disks to permit 
multiple file changes without affecting users, but they must be followed 
very carefully.  Better to not fool Mother Nature with any production 
minidisks in the first place until you have researched more, and as 
always: have good backups in place and ready for restore.

Mike Walter
Hewitt Associates
The opinions expressed herein are mine alone, not my employer's.



"Michael Harding" <[email protected]> 

Sent by: "The IBM z/VM Operating System" <[email protected]>
12/05/2009 02:10 PM
Please respond to
"The IBM z/VM Operating System" <[email protected]>



To
[email protected]
cc

Subject
Re: Automate z/VM RACF SMF process to z/OS






The in-core directory is why my first step was to tell racfvm to release 
its a-disk. That flushes the in-core copy. Although I believe when you 
tell it to access the disk after you've done the flip/change/flip sequence 
the in-core copy is discarded, but don't offhand remember that for sure.
--
Mike Harding
z/VM System Support

[email protected]
[email protected]
[email protected]
(925) 926-3179 (w)
(925) 457-9183 (c)
IM: VMBearDad (AIM), mbhcpcvt (Y!)


The IBM z/VM Operating System <[email protected]> wrote on 
12/05/2009 03:01:14 AM:

> From: Kris Buelens <[email protected]>
> To: [email protected]
> Date: 12/05/2009 03:02 AM
> Subject: Re: Automate z/VM RACF SMF process to z/OS
> Sent by: The IBM z/VM Operating System <[email protected]>
> 
> I just made a test, CMS is still behaving badly: it ignores the 
> machine check or CP isn't giving one when a R/W disk gets linked R/
> O.  Conclusion: simply using CP SEND LINK ... RR followed by an 
> update and CP SEND LINK .... M is as dangerous as doing a LINK MW.
> The reason: CMS has structures in storage describing an accessed 
> minidisk, when you suddenly LINK a R/W mdisk in R/O, CMS doesn't do 
> anything, the later R/W link doesn't trigger anything else in CMS 
> either, so the in storage structure still describes the initial 
> state of the minidisk.  Changes applied by another user are invisible.




The information contained in this e-mail and any accompanying documents may 
contain information that is confidential or otherwise protected from 
disclosure. If you are not the intended recipient of this message, or if this 
message has been addressed to you in error, please immediately alert the sender 
by reply e-mail and then delete this message, including any attachments. Any 
dissemination, distribution or other use of the contents of this message by 
anyone other than the intended recipient is strictly prohibited. All messages 
sent to and from this e-mail address may be monitored as permitted by 
applicable law and regulations to ensure compliance with our internal policies 
and to protect our business. E-mails are not secure and cannot be guaranteed to 
be error free as they can be intercepted, amended, lost or destroyed, or 
contain viruses. You are deemed to have accepted these risks if you communicate 
with us by e-mail. 

Reply via email to