Richard,

I'm not sure exactly what you mean by "sharing," in this case.  Is it simply
having the same DASD in the I/O gen for each LPAR, as opposed to trying to
actually share the data?  If the latter, don't do that, there's no means to
insure data integrity.  If the former, I would still advise against it.
There have been problems in the past with table overflows for particular
device types when there were too many devices online to the system.

Having the DASD for the Linux/390 system specified via the "dasd=" parameter
doesn't guarantee that someone won't use the dynamic capabilities of the 2.4
kernel to bring a volume online and format it.  Oops, there goes the payroll
database.

In general not a good idea all the way around.

Mark Post

-----Original Message-----
From: Richard W. Lauck, Cornerstone Systems, Inc.
[mailto:[EMAIL PROTECTED]
Sent: Wednesday, March 05, 2003 1:45 PM
To: [EMAIL PROTECTED]
Subject: S/390 RedHat Shared DASD


How many are sharing DASD between Linux and OS LPARs rather than having a
few volumes dedicated to the Linux LPAR?.    What effects of the Linux
autosensing all shared devices and causing sharing problems that IBM warns
about have you seen?   Is just having something like

dasd=192-194,200

in the parmfile enough even though all DASD is shared between LPARs?








Richard W. Lauck
Cornerstone Systems, Inc.
Sr. Systems Programmer
IBM Certified S/390 Parallel Sysplex Systems Programmer
IBM Certified S/390 Parallel Sysplex Operator
IBM Parallel Sysplex Top Gun
(425)489-4579                     Direct - Home Office
(425)453-5166  x9024        Voice Mail
(425)486-4501                     Home
(888)505-4534                     Pager

Reply via email to