Re: Tape; 2nd port of 5735; 8 Gigabit PCI Express Dual Port Fibre Channel Adapter

You can in fact have an IBM i Host with Fiber with drives on the fiber and share those drives with one OR MORE guest partitions. It all works very well. You will note that I said "Drives" and I specifically did NOT say "Libraries." Even drives that are part of libraries cannot be shared if i is the host.

- Larry "DrFranken" Bolhuis

www.Frankeni.com
www.iDevCloud.com - Personal Development IBM i timeshare service.
www.iInTheCloud.com - Commercial IBM i Cloud Hosting.

On 12/21/2015 2:26 PM, Steinmetz, Paul wrote:

Am I correct that Fiber Adapters can only be virtualized if using VIOS?
Not capable with I hosting I, correct?

Paul

-----Original Message-----
From: MIDRANGE-L [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of DrFranken
Sent: Monday, December 21, 2015 2:20 PM
To: Midrange Systems Technical Discussion
Subject: Re: Tape; 2nd port of 5735; 8 Gigabit PCI Express Dual Port Fibre Channel Adapter

oops. :-)

Yes it did in fact replace. You may ONLY have one map per virtual fiber adapter.

Then depending on the zoning in your switch that you have now just connected to you could 'see' many things you previously couldn't.

What *I* would do to fix this is to remove the mapping, I think you merely do:

fcmap -vadapter vfchost# -fcp

Since you omitted the port it maps it Not. Of course you can use the HMC in Virtual Storage management Fiber Chanel tab and uncheck it and apply that. You will need to check the 'remove from active partition'
checkbox.

Then I would DLPAR away the virtual fiber adapter.

Then in SST clear all non-reporting stuff.

Then DLPAR the bugger back in.

Then re-map it.

I seem to remember Jim O doing one of these things and ending with with every tape drive showing up on every LPAR in his lab about 8 times. :-)


- Larry "DrFranken" Bolhuis

www.Frankeni.com
www.iDevCloud.com - Personal Development IBM i timeshare service.
www.iInTheCloud.com - Commercial IBM i Cloud Hosting.

On 12/21/2015 2:02 PM, rob@xxxxxxxxx wrote:

I had successfully set up one tape library (3573-L4U) to be shared via
multiple lpars.
I set up the virtual between the VIOS lpar with the card to each client
lpar.
Ran the
vfcmap -vadapter vfchost# -fcp fcs0
where # is tied to each client lpar. Fully realizing that vfchost2 does
not necessarily mean lpar 2.
Works great.
From that 5735 card it goes to a san switch. On the IBM i you see
WRKHDWRSC *STG
Resource Type-model Status Text
CMB03 6B25-001 Operational Storage Controller
DC03 6B25-001 Operational Storage Controller
TAPGRT01 3573-040 Operational Tape Library

Option 9 on the media library shows:
Resource Type-model Status Text
TAPGRT01 3573-040 Operational Tape Library
TAP02 3580-004 Operational Tape Unit
TAP01 3580-004 Operational Tape Unit
TAP03 3580-004 Operational Tape Unit
TAP04 3580-004 Operational Tape Unit

WRKMLBSTS shows
Device/
Resource Status Allocation
TAPGRT01 VARIED ON
TAP01 OPERATIONAL UNPROTECTED
TAP04 OPERATIONAL UNPROTECTED
TAP02 OPERATIONAL UNPROTECTED
TAP03 OPERATIONAL UNPROTECTED

When initially setting this up a long time ago I had ran:
chgdev -dev fscsi0 -attr fc_err_recov=fast_fail dyntrk=yes -perm
chgdev -dev fscsi1 -attr fc_err_recov=fast_fail dyntrk=yes -perm
and rebooted vios (before all the config above).

All working good. BRMS, etc. Been this way for over a year.

Now I thought, I'd like to hook up that second port on that card to
another library.
Ran:
vfcmap -vadapter vfchost2 -fcp fcs1

All heck breaks loose. On a different lpar than the WRKHDWRSC *STG above,
I now see:
Resource Type-model Status Text
CMB03 6B25-001 Inoperative Storage Controller
DC03 6B25-001 Inoperative Storage Controller
TAPGRT01 3573-040 Inoperative Tape Library

WRKMLBSTS was wonky so I varied it off but now I cannot vary it on. Which
kind of makes sense since the hardware is "Inoperative".

My first thought is how do I back out the vfcmap command?

I ran a
lsmap -all -npiv | more
and I see:

Name Physloc ClntID ClntName ClntOS
------------- ---------------------------------- ------ --------------
-------
vfchost1 U8286.41A.214405V-V2-C204 4 MAIL3 IBM
i

Status:LOGGED_IN
FC name:fcs0 FC loc code:U78C9.001.WZS00U9-P1-C9-T1
Ports logged in:5
Flags:a<LOGGED_IN,STRIP_MERGE>
VFC client name: VFC client DRC:U8286.41A.214405V-V4-C204

Name Physloc ClntID ClntName ClntOS
------------- ---------------------------------- ------ --------------
-------
vfchost2 U8286.41A.214405V-V2-C205 5

Status:NOT_LOGGED_IN
FC name:fcs1 FC loc code:U78C9.001.WZS00U9-P1-C9-T2
Ports logged in:0
Flags:4<NOT_LOGGED>
VFC client name: VFC client DRC:


Judging by this it didn't ADD fcs1. It REPLACED fcs0 with fcs1.

So, to back it out, let me try
vfcmap -vadapter vfchost2 -fcp fcs0

lsmap shows fcs0 now but it still isn't "logged in". WRKHDWRSC *STG still
shows it as inoperative. Sandbox lpar so I am thinking IPL. In
progress...
Meanwhile I am wondering, since it 'replaced' instead of added, if I need
to configure a separate virtual fiber channel between this vios lpar, and
the IBM i client lpar and then use that new vfchost# in vfmap...

Rob Berendt