Category: VIO

Configure Virtual Tape on VIO Server ( Map physical tape to AIX Client)

You may need to backup AIX client data to tape, the VIO allow mapping physical tapes starting with VIOS v2.1.0.10-FP20.1.

But it has some limitation on the type of tapes that can be mapped (Check IBM documentation)

on the VIO server you can check wether the tape can be mapped by issuing the following command

VIO_SERVER$lsdev -type tape

this will list the available tapes on the system.

VIO_SERVER$lsdev -type tape4vtd

this will give you the list of tape that can be maped.

VIO_SERVER$mkvdev -vdev rmt0 -vadapater vhost0

vhost0 is the VIRTUAL SCSI adapter mapped to VIO Client LPAR

Easy way to get Virtual SCSI mapping from VIO Client

Sometimes you need to get the VIO mapping from the client side.

the easiest way is to run the below command:

root@client_lpar: / # echo “cvai” | kdb | grep vscsi

read vscsi_scsi_ptrs OK, ptr = 0xF1000000C01A83C0
vscsi0 0x000007 0x0000000000 0x0 VIO_SRV1->vhost0

VIO NPIV Mapping to FC / Unmap

In this post i will share with you some tips to configure NPIV on VIO. NPIV is very usefull on VIO and can reduce overhead for administering storage on the VIO. Opposed to virtual SCSI, NPIV help storage admin to have full visibility on the LPARs storage disks.

First check fc ports on vio servers that support NPIV using lsnports command. This command also displays the remaining number of WWN that can be assigned by this FC adapter.

$ lsnports
name             physloc                        fabric tports aports swwpns awwpns
fcs0             U5803.001.91800ZL-P2-C3-T1          1     64     64   2048    2047
fcs1             U5803.001.91800ZL-P2-C3-T2          1     64     63   2048    2046
fcs2             U5803.001.91800ZL-P2-C5-T1          1     64     64   2048    2047
fcs3             U5803.001.91800ZL-P2-C5-T2          1     64     64   2048    2047
fcs4             U5803.001.91800ZL-P2-C6-T1          1     64     64   2048    2047

To map virtual FC adapter vfchost4 to Physical FC adapter fcs1 type:
$vfcmap -vadapter vfchost4 -fcp fcs1

Know check if the mapping was successful, type:
$lsmap -all -npiv | more
Name Physloc ClntID ClntName ClntOS
————- ———————————- —— ————– ——-
vfchost4 U9119.FHA.44337FA-V2-C165 160 WASUAT AIX

Status:LOGGED_IN
FC name:fcs1 FC loc code:U5803.001.91800ZL-P2-C3-T2
Ports logged in:1
Flags:a<LOGGED_IN,STRIP_MERGE>
VFC client name:fcs0 VFC client DRC:U9119.FHA.44337FA-V160-C5

In some cases you may want to remove the connection between Virtual FC adapter and the physical adapter, type
$ vfcmap -vadapter vfchost4 -fcp

Know check again that it was unmapped.

$lsmap -all -npiv | more
Name Physloc ClntID ClntName ClntOS
————- ———————————- —— ————– ——-
vfchost4 U9119.FHA.44337FA-V2-C165 160

Status:NOT_LOGGED_IN
FC name: FC loc code:
Ports logged in:0
Flags:1<NOT_MAPPED,NOT_CONNECTED>
VFC client name: VFC client DRC:

For more info about tips for implementing NPIV on VIO, check this article on IBM Developer Works.

http://www.ibm.com/developerworks/aix/library/au-NPIV/

Enable High Bandwith communication between LPARs on VIO

By default virtual adapters can go up to 1.3 Gbps. If you need to boost the network performance between LPARs we have 2 options:

1- MTU Size larger than 1500

2- AIX largesend option

If the LPAR is communicating with external Servers, Jumbo frames should be enabled end-to-end in order to have the full benefit. otherwise packets will be fragmented. Inside VIO we don’t have this issue since we are using the internal virtual switch. So we can even go up to higher MTU size, the maximum is 65280.

So here is a procedure to check the network performance before and after setting MTU size.

# lsattr -El en3 | grep “^mtu”

mtu 1500 Maximum IP Packet Size for This Device True

#ftp> put “|dd if=/dev/zero bs=64k count=100000” /dev/null

write down the result we got : 1.258e+05 this means 1.2580×100000 Kbytes/s

Change the MTU Size:

#chdev -l en3 -a mtu=65280

Redo the ftp test

#ftp> put “|dd if=/dev/zero bs=64k count=100000” /dev/null

the result was 3.025e+05 which gave us an increase in performance by 60 %

No RMC Connection when trying to add a new virtual adapter to an LPAR using HMC

I had this issue today, when i try to add a virtual adapter to different logical partitions. “No RMC Connection”

In our case the issue was caused due to the AIX image created using Original clone hosted on another HMC.

Here are the steps to recreate the configuration and reconnect the LPAR to the HMC.
1- check rmc connection status from the LPAR
hostname:/# cd /usr/sbin/rsct/bin
hostname:/usr/sbin/rsct/bin# ./rmcdomainstatus -s ctrmc

Management Domain Status: Management Control Points
O A 0xed5304fc5aaf4012 0001 10.99.27.101
O Indicates that the RMC connection is “Down”, as determined by the RMC heartbeat mechanism.

2- list rsct services status
hostname:/usr/sbin/rsct/bin# lssrc -a | grep rsct
ctrmc rsct 3211382 active
IBM.DRM rsct_rm 7077892 active
IBM.CSMAgentRM rsct_rm 7012524 active
IBM.ServiceRM rsct_rm 6422688 active
ctcas rsct inoperative
IBM.HostRM rsct_rm inoperative
IBM.ERRM rsct_rm inoperative
IBM.MgmtDomainRM rsct_rm inoperative
IBM.AuditRM rsct_rm inoperative
hostname:/usr/sbin/rsct/bin#

You will see some active and some missing (The key for DLPAR is the IBM.DRM)

3- recreate the configuration
hostname:/usr/sbin/rsct/install# cd bin
hostname:/usr/sbin/rsct/install/bin# ./recfgct
0513-071 The ctcas Subsystem has been added.
0513-071 The ctrmc Subsystem has been added.
0513-059 The ctrmc Subsystem has been started. Subsystem PID is 3211402.
hostname:/usr/sbin/rsct/install/bin#

4 – restart services for rsct
hostname:/usr/sbin/rsct/install/bin# cd ../..
hostname:/usr/sbin/rsct# cd bin
hostname:/usr/sbin/rsct/bin# rmcctrl -z
hostname:/usr/sbin/rsct/bin# rmcctrl -A
0513-071 The ctrmc Subsystem has been added.
0513-059 The ctrmc Subsystem has been started. Subsystem PID is 5832812.
hostname:/usr/sbin/rsct/bin# rmcctrl -p

5 – recheck rmc domain status
hostname:/usr/sbin/rsct/bin# ./rmcdomainstatus -s ctrmc
hostname:/usr/sbin/rsct/bin#

it’s normal if you don’t get any result.
wait some time and re-issue the command.

7- hostname:/usr/sbin/rsct/bin# r ./
./rmcdomainstatus -s ctrmc

Management Domain Status: Management Control Points
I A 0xed5304fc5aaf4012 0001 10.99.27.101

Now the status is I which indicates that we have an active connection with the HMC.

IBM VIO Server and VMAX device ID

Sometimes we need to check the device ID between VIO Server and VMAX.

there is multiple methods to locate the required device. the two methods that i use always.

from VIO shell $ chkdev -dev hdsik1

form aix shell #odmget -q “name like hdisk1 and attribute=unique_id” CuAt

the value returned should contain the VMAX serial number followed by the device ID