Moving the DVD-RAM Between LPARs using the VIO server

Technote (FAQ)

Question
Is there a method to share the DVD-RAM or DVD-ROM device on System p servers that does not require a dynamic
logical partition (DLPAR) action?
Cause
Limitations of the hardware
Answer
One of the common features normally available with HMC attached System p servers is the ability to use Dynamic
Logical Partitioning (DLPAR) to add/move/remove I/O devices such as the CDROM or DVD controller between LPARs
without taking an outage. The functionality required for DLPAR actions includes an active network connections
between the LPARs and the HMC over port 657. If you have a virtual I/O (VIO) server partition that owns the
DVD controller then you can run a few simple commands on the VIO server to allow you to map cd0 from one
client LPAR to another using virtual SCSI.

The process to move cd0 from one logical partition to another using VIO server commands is illustrated below.
The example assumes the user is logged in as padmin on VIO server:

- Determine if the VIO server owns an optical device we use the lsdev command.

$ lsdev -type optical

name status description
cd0 Available IDE DVD-ROM Drive

- Determine if cd0 is already mapped to a client LPAR we use the lsmap command.

$ lsmap -all |more

SVSA Physloc Client Partition ID
--------------- -------------------------------------------- ------------------
vhost0 U9111.520.104797E-V1-C11 0x00000002

VTD vtscsi0
LUN 0x8100000000000000
Backing device rootvg2a
Physloc

SVSA Physloc Client Partition ID

--------------- -------------------------------------------- ------------------
vhost1 U9111.520.104797E-V1-C13 0x00000003

VTD vtscsi1
LUN 0x8100000000000000
Backing device rootvg3a
Physloc

VTD vtscsi2
LUN 0x8200000000000000
Backing device datavg3a
Physloc

- Looking through the "Backing device" section of each vhosts we do not see cd0 listed. We could have run
"lsmap -all | grep cd0" as a quick check as well.

- To assign device cd0 to LPAR ID 2 we first need to locate its associated virtual SCSI server (vhost) device
from the output listed above. If you look at the "Client Partition ID" section of the lsmap output you can
see that vhost0 is associated with partition ID 2 (hex 0x00000002).
- To make the virtual SCSI map of cd0 to LPAR ID 2, we use mkvdev as follows:

$ mkvdev -vdev cd0 -vadapter vhost0

vtopt0 Available

- To check to see if cd0 and vtopt0 show up under vhost0 resources we use lsmap.

$ lsmap -vadapter vhost0

SVSA Physloc Client Partition ID
--------------- -------------------------------------------- ------------------
vhost0 U9111.520.104797E-V1-C11 0x00000002

VTD vtopt0
LUN 0x8200000000000000
Backing device cd0
Physloc U787A.001.DPM06E2-P4-D2

VTD vtscsi0
LUN 0x8100000000000000
Backing device rootvg2a
Physloc

By virtually mapping the DVD device cd0 to LPAR 2's vhost, there is no need to make changes to the LPAR's
profile or perform further actions such as DLPAR functions from the HMC. The cd0 device is now ready for the
client LPAR to use. If the LPAR is already in a running state, the cfgmgr command would need to be run as
root on the client LPAR so the new device could be configured. If the LPAR is not activated then once it is
started, the DVD device will be available for performing installation or maintenance functions on the LPAR.

Once LPAR 2 is finished using the DVD, it can then be removed and mapped to a different LPAR if desired. For
example, suppose that LPAR ID 3 needed the DVD for maintenance, by removing the VTD name from vhost0 and
making a new virtual SCSI map to vhost1, we would be giving LPAR 3 access to the DVD. Following commands
illustrate the actions required.

$ rmdev -dev vtopt0 -recursive

vtopt0 deleted

$ mkvdev -vdev cd0 -vadapter vhost1

vtopt0 Available

$ lsmap -vadapter vhost1

SVSA Physloc Client Partition ID

--------------- -------------------------------------------- ------------------
vhost1 U9111.520.104797E-V1-C13 0x00000003

VTD vtopt0
LUN 0x8500000000000000
Backing device cd0
Physloc U787A.001.DPM06E2-P4-D2

VTD vtscsi1
LUN 0x8100000000000000
Backing device rootvg3a
Physloc

VTD vtscsi2
LUN 0x8200000000000000
Backing device datavg3a
Physloc

The process of virtual mapping of the optical device, cd0, between client LPARs of a VIO server is much
simpler than trying to perform DLPAR related functions to achieve the same result since all the interaction
takes place on the VIO server command line.

==============================================================================

Moving a DVD or CD drive to another LPAR

Technote (FAQ)

Question
How can I move my DVD-ROM or CD-ROM drive from one LPAR to another?
Answer
If you don't know which LPAR owns the CD-ROM drive, use the HMC manager or WEBSM tool.

Select the managed system and open "Properties".

Select the "I/O" tab. Look for the I/O device with the description "Other Mass Storage Controller" and read
the "Owner" field. This will show the LPAR currently owning that device.
ON THE SOURCE SYSTEM
1. Find the parent adapter of the DVD or CD device:

$ lsdev -Cl cd0 -F parent

ide0

2. Find the slot containing the IDE bus:

$ lsslot -c slot
# Slot Description Device(s)
U787B.001.DNWG2AB-P1-T16 Logical I/O Slot pci1 ide0
U9133.55A.105C2EH-V7-C0 Virtual I/O Slot vsa0
U9133.55A.105C2EH-V7-C2 Virtual I/O Slot ent0
U9133.55A.105C2EH-V7-C3 Virtual I/O Slot vscsi0

so in this example PCI1 is the slot containing the IDE adapter and CD drive.

3. Remove the slot from this host:

# rmdev -dl pci1 -R

cd0 deleted
ide0 deleted
pci1 deleted

ON THE HMC

Select the LPAR currently owning the CD-ROM, and in the Actions menu select:
Dynamic Logical Partitioning -> Physical Adapters -> Move or Remove
Select the adapter for "Other Mass Storage Controller" and move to the desired target LPAR.
This will perform a DLPAR operation on both the source and target LPAR.

ON THE TARGET SYSTEM

Log in as root and run

# cfgmgr

The CD-ROM device should show up now

# lsdev -C | grep cd
cd0 Available 1G-19-00 IDE DVD-ROM Drive

==============================================================================

Shared DVD For lpars

 发布日期: 2015 年 9 月 17 日

Hossam Moghazy
关注 Hossam Moghazy
Senior Infrastructure Consultant Linux/AIX at Huawei Technologies
Hi ;
issues may face us during creating lpars and installing aix on it is to share lpar
here it is the steps in snapshots below starting from the main screen above
the point is to make sure that resource "DVD" shared as read only
then mark on force assign check button
thanks
Hossam Moghazy
My Hands-On Experience With POWER8
August 2014 | by Jaqui Lynch

Print <=""> 0
References
In June, my company received its first two POWER8 servers to use for testing. One server was a single-socket S814 with
PowerVM, RHEL and AIX, and the second was a Linux-only box running PowerKVM and RHEL Linux. I had the privilege of
working with the AIX server, which I attached to an HMC and our V7000 disk subsystem.
The Specs
The S814 is configured with the high-function backplane so it wasn’t a split backplane. It has 8 x 3.7ghz cores, 128GB of
memory, 12 x 387GB SSDs, 4 x 300GB hard drives, 2 x dual port 16Gb Fibre cards, 2 x 4 port dual 10Gb and dual 1Gb FCoE
cards as well as the 4 port 1Gb network card. I chose the FCoE cards rather than the new 10Gb network cards because they’re
supported by network install manager (NIM). The S822L Linux only box has 10 x 3.4ghz cores and 64GB of memory with a 4
port 1Gbe adapter and 2 x 300GB hard drives. Because the S822L is running PowerKVM, it’s not connected to an HMC. This
article will focus on the S814.
The S814’s high-function backplane means all disks and SSDs were on the same controller, which went to one VIO server. The
second VIO server will be booted from SAN and configured later. I installed two client LPARs—one running AIX v7.1 tl03 sp3
and the other running AIX v6.1 tl09 sp3. The HMC was a CR7 and was installed at the latest 8.8.1.0 MH01441, which is still the
highest level.
The Setup
After connecting everything, powering up the server and flashing the microcode, I ensured the logical memory block or memory
region size was set to the same as our other servers. This is required using live partition mobility and you must power cycle the
box if you have to change it. The configuration and setup of the VIO and LPAR profiles was no different to any other server and
very few differences were noted in the HMC GUI, so the setup was simple. Since the NIM server had been upgraded to AIX
v7.1 tl03 sp3, the next step was to create the LPP source and SPOT for NIM for the new AIX v6 and v7 levels to be installed.
Additionally, I used installios against the VIO server DVDs to create the necessary VIO resources on the NIM master.
The VIO images were also loaded to an FTP location on the NIM server. After creating the VIO profile, I booted the LPAR and
checked the Yes box at the top of the activation pop-up that says, “Install Virtual I/O Server as part of activation process?” Once
the VIO server was running, I went through all the normal checks. Ioslevel shows as 2.2.3.3 and oslevel –s shows the operating
system at 6100-09-03-1415. This means the VIO will be running in SMT4 mode since SMT8 requires 7.1 tl03 sp3. The following
shows this is the case:
# smtctl
This system is SMT capable.
This system supports up to 4 SMT threads per processor.
SMT is currently enabled.
After setting up the VIO on two of the four internal disks, I then set up FBO (file backed optical) and loaded a bunch of ISO
images for various sets of software. I always set up FBO if I have a spare disk as I tend to lose the DVDs. By ripping them to
ISO images and uploading them to FBO, I can always find them and recreate the DVDs as needed. And, of course, I can load
them remotely to any of the LPARs controlled by that VIO server. After setting up FBO, it looked as follows:
$ lssp
Pool Size(mb) Free(mb) Alloc Size(mb) BDs Type
rootvg 540672 407552 512 2 LVPOOL
fbovg 1089024 577024 512 1 LVPOOL

$ lsrep
Size(mb) Free(mb) Parent Pool Parent Size Parent Free
511414 490594 fbovg 1089024 577024
From this, we can tell that the FBO pool is approximately 1TB in size with about half of that space unused.
$ lsvopt
VTD Media Size(mb)
vtopt0 No Media n/a
vtopt1 No Media n/a
vtopt2 No Media n/a
vtopt3 No Media n/a
Finally, with the VIO server up and running it was time to create the client LPARs. I created an AIX v7 LPAR and an AIX v6
LPAR, both with an entitlement of three and six virtual processors (VPs) and 32GB of memory. Both were provided SAN-based
disks via vSCSI from the VIO server. Once everything was running, I ran smtctl to check the LPARs. It turns out that by default
the AIX v7 LPAR comes up in SMT4 initially. I ran ‘smtctl –t 8’ to change the AIX v7 LPAR to SMT8. vmstat shows an
entitlement of 3 and logical CPUs (LCPUs) of 48 on the AIX v7 LPAR and LCPUs of 24 on the AIX v6 LPAR.

SMT8 Support and Performance

To get SMT8 support on POWER8, the LPAR needs to run AIX v7.1 TL3 SP3, otherwise the LPAR will run in SMT4 mode. I
decided to use nmem, ncpu and my own test program written in C to compare the three LPARs, which all had the same
entitlement, VPs and memory. On the S814, one LPAR ran AIX v7 with SMT8 and a second ran AIX v6 with SMT4. The third
ran AIX v7 with SMT4 on a 750B (which has 32 x 3.3ghz cores). Memory on the 750 was 1066mhz and on the S822L and S814
was 1600mhz, which puts the 750 memory at about 66 percent of the speed of the S814 memory. Comparing rPerfs for three
cores on each shows the 750 LPAR to be estimated at 60.42 percent of the rPerf of the SMT8 AIXv7 POWER8 LPAR and 64.66
percent of the SMT4 AIXv6 POWER8 LPAR. The AIX v6 SMT4 POWER8 LPAR is estimated at 93.44 percent of the AIX v7
SMT8 POWER8 LPAR. So how did they stack up?

Using nmem I saw the following as an overall average (OPS/S K is thousands of operations per second):
OPS/S K

b814aix1 29820.78 P8 AIX v7 SMT8

b814aix2 27483.75 P8 AIX v6 SMT4

b850nl1 21708.83 P7 AIX v7 SMT4

 Comparatives OPS/S K

nl1/aix1 73.89%

aix1/nl1 135.54%

nl1/aix2 81.94%

aix2/nl1 122.69%

aix2/aix1 90.43%

aix1/aix2 110.78%

Comparing the OPS/S K, we can see the 750 LPAR performed better than predicted by the difference in mhz in the memory. As
we all know, mhz is not everything; variables like buffering, cache sizes and even instruction type all impact how quickly things
move through memory. With that said, the memory performance in operations per second was still 22 percent to 35 percent
better on POWER8 than it was on POWER7, which is a significant improvement.
Comparatives rPerf Results (128 processes) C Program

nl1/aix1 60.42% 67.95% 59.43%

aix1/nl1 165.5% 147.17% 168.28%

nl1/aix2 64.66% 74.19% 71.31%

aix2/nl1 154.65% 134.78% 140.23%

aix2/aix1 93.44% 91.58% 83.33%

aix1/aix2 107.02% 109.19% 120%

When comparing rPerf, I’d expect the 750 LPAR to be between 60 and 65 percent of the POWER8 LPARs and the AIXv7
POWER8 LPAR to be around 7 to 8 percent faster than the AIX v6 POWER8 LPAR. Multiple tests were run with ncpu using
varying numbers of processes. The results reported are averages from multiple runs using 128 processes. In all three LPARs,
this exceeded the number of threads available to the LPAR and it stressed the CPU accordingly. When comparing the user time
from ncpu for 128 processes on each LPAR to the other LPARs, it’s clear that the two POWER8 LPARs seem to scale very
closely to how the rPerf scales. The POWER7 LPAR did better than expected using 47 percent more user time than the AIX v7
P8 LPAR instead of 65.5 percent more and 34.78 percent more user time than the AIX v6 P8 LPAR instead of 54.65 percent.
Similar results were seen using the C program written to drive CPU.

My Conclusions

I should note that these were very limited tests and there are many more tests left for me to run in the benchmarking suites that
I am testing. Additionally, they don’t test all of the functions of the server and are not a true mixed OLTP workload, which is what
most run. However, they do provide some initial data that shows that POWER8 appears to scale as expected, both due to
improvements in the memory performance as well as CPU performance. Also the jump from SMT4 to SMT8 provides around a
9 percent boost, which is on a par with what’s predicted in the published rPerf. Further, more detailed tests are planned using
several test suites for memory, cpu, I/O and network performance. Although preliminary and limited, they provide a window into
the performance potential of the new POWER8 scale-out servers.

Overall, the POWER8 experience has been very positive so far. Firmware and HMC updates went smoothly and the server
appears to be performing as expected. This provides a level of confidence that you can move from POWER7 to POWER8 while
reducing the VPs or cores in LPARs using rPerf comparisons as an approximate scaling factor.
 1
 2

Jaqui Lynch is an independent consultant, focusing on enterprise architecture, performance and delivery on Power Systems with AIX and Linux.

More Articles From Jaqui Lynch

Like what you just read? To receive technical tips and articles directly in your inbox twice per month, sign up for the EXTRA e-
newsletter here.