In my experience, the majority of the devices a Linux guest finds, at least
the ones not needed, are DASD devices.
So set up a single minidisk that all servers link to RR as their 191 disk.  

The common PROFILE EXEC can then be tuned to perform work common to all
servers (e.g. CP DETACH any unneeded MDISKs), and also perform special setup
for special servers, and then IPL from the proper Linux server boot disk.

With one PROFILE EXEC for all servers, site-wide changes become consistent,
and easy (as long as the change was tested on a different server first, and
then renaming the current PROFILE EXEC on the common 191 disk to something
like "-1PROFILE EXEC"), and copying the updated PROFILE EXEC to the common
191 disk.

Mike Walter

-----Original Message-----
From: Linux on 390 Port [mailto:LINUX-390@VM.MARIST.EDU] On Behalf Of Mark
Post
Sent: Monday, January 30, 2017 2:15 PM
To: LINUX-390@VM.MARIST.EDU
Subject: Re: Unable to IPL the system after resizing "/" btrfs filesystem

>>> On 1/30/2017 at 09:19 AM, Michael J Nash <miken...@us.ibm.com> wrote: 

> Greetings Mark, thank you for your reply.
> Please explain the reasoning for disabling cio_ignore.
> The reason I would not want to do this is because these guest systems 
> may be used for IPL of different distributions and/or releases.
> I may IPL a SUSE 12, Red Hat 6.7, or a Red Hat 7.2 system.  Having 
> disks visible used by other systems may cause problems.
> Also, LVM volumes may conflict with naming conventions, For example: 
> both systems using the same volume group name.

The cio_ignore feature was introduced primarily for LPAR installations.  It
helps speed up booting because the kernel doesn't try to enumerate every
single device it can find.  For z/VM environments that have relatively very
few devices available, it's not needed.  I can't speak for RHEL systems, but
I think it's pretty much the same as for SLES.  That is, unless you make a
device persistently online, it's not going to come online by itself, so you
shouldn't run into problems with LVM, extraneous DASD volumes being online,
etc.  Disabling cio_ignore in a guest means that people won't be constantly
tripping over it when it's something they've not had to do for over 15
years.  There are a number of things that you have to remember to do
yourself when cio_ignore is active if you're not using YaST to do things.
In general, it's a PITA for guests.


Mark Post

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions, send email
to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For more information on Linux on System z, visit http://wiki.linuxvm.org/

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to lists...@vm.marist.edu with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For more information on Linux on System z, visit
http://wiki.linuxvm.org/

Reply via email to