In theory yes you could that would set the PTF images in one place which could be used by many partitions. However you first must be careful that you're getting all the PTFs because by default SNDPTFORD won't download PTFs you already have. THus attempting to use these on another system or partition may fail with unsatisfied dependencies. Another issue is letting the system name the files you don't know when you can delete them.

You tried what I first tried and of course itnoworkie. The problem here is that you must have the images in a place where the image catalog can properly access them EVEN IF it's in restricted condition. (e.g. where all communications is shut down.)

So what I do instead is to use Fix Central to download each group as it changes into the correct directory for that release. Such as /ibmi/v7r2/ptf. The files are named by their group and number so DB19 for the database group level 19. When DB20 comes out I order that and can now delete the DB19 files. Because I do NOT tell fix central to query my system for existing PTFs it creates full groups with every pre and co req. Yes they are bigger but I only need one set this way.

Now how does one access this set, as Rob's way fails. Fortunately way back in the annals of unsupported releases IBM had already included support for such a thing and it's really pretty cool and trivially simple to the most casual of observers.

1) Order and download the images into the shared place as I describe above.
2) Create the image catalog on that system as you would for any system.
3) When you verify it (you ALWAYS verify it first, right?) enter *YES in the NFSSHARE Parm. This creates the file VOLUME_LIST in the image catalog directory and populates it.
4) Assure the VOLUME_LIST file has CCSID 819. Failure to do this may drive you batty. e.g.:

CHGATR '/ibmi/v7r2/ptf/VOLUME_LIST' ATR(*CCSID) VALUE(819)

5) Gve the public read and execute authority to the directory tree.


6) Share the directory read only much like Rob did below. e.g.

CHGNFSEXP OPTIONS('-i -o ro') DIR('/ibmi/v7r2/ptf')

The PTFs are now available to any system but you do need to set up access to them and that's where Rob and I failed initially. We used straight up NFS. (Actually I also tried /QFileSvr.400 and /QNTC as well. I'm a slow learner)

This requires two bits to accomplish:

1) Assign a service tools interface to the partition. That interface must be on the same network or be able to reach the network that contains the share above. This is done in Service Tools (STRSST) and then option 8 then F13. Select an available Ethernet port (one that is NOT being used by IBM i and yes it may be a virtual interface) and assign it an IP address and gateway. Save it and remember to activate it as well.

Ignore the warning that "Selected resource is not full LAN console capable" as you're not using the interface for that. Side note: If you already have LAN console on this partition then this step is already completed!

Note that you MUST Be able to PING from the NFS Host partition to this service tools address. Until that works DO NOT proceed as you will not get far.

2) Create a network based virtual optical drive.
TEXT('PTF Images on NFS Hosting Partition')

Where: is the IP address of the server where the CHGNFXEXP was performed.
'/ibmi/v7r2/ptf' matches *EXACTLY to the export. Case Is SeNSitIve!

Now vary on the virtual optical and if all is well then WRKOPTVOL will show the first image in the catalog as available. Use this in GO PTF option 8 or INSPTF. It will roll through all the images in the VOLUME_LIST file.

A few notes.
1) After use VARY OFF the OPTNET01 device. This is important as it only queries the VOLUME_LIST file at Vary ON time so you'll want that to happen when you need it.
2) No Jumbo Frames. Itnoworkie.
3) The Image Catalog on the NFS host is not used. It's only use is to assist you in validating the images are all there and valid and to create the VOLUME_LIST file. You do not need it to be mounted in a virtual optical drive to use it.
4) Performance is vastly superior this way than copying all those images to a partition and using them there. First you save the copy and second the disk space and third the I/O load is split - reading on the host and updating IBM i on the target.
5) Any number of partitions can access these images at the same time. Each is only reading via NFS the various image files.
6) If you DO replace DB19 with DB20 you do not need to redo the entire image catalog on the host. Simply edit the VOLUME_LIST file and change DB19_1.bin to DB20_1.bin etc and you are good to go.

No Rob you can't write to such an image catalog, tried it, got the raspberries. Read Only. Way ahead of you there.... :-(

And there you have it piece of cake.

Merry Christmas!!

- Larry "DrFranken" Bolhuis

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

On 12/23/2015 1:50 PM, rob@xxxxxxxxx wrote:
On a new lpar,
which happens to have a fully loaded
I ran

On a different lpar I ran:
MOUNT TYPE(*NFS) MFS('domtest:/fixes') MNTOVRDIR('/domtest/fixes')

I can do
WRKLNK '/domtest/fixes/cume/*'
and see stuff. But when I try
I get
CPD4F06 - Unable to determine absolute path name.
Recovery . . . : Make sure the directory resides in a file system
by the image catalog commands.

I'm suspecting that mounts are not supported...

Rob Berendt

This thread ...


Follow On AppleNews
Return to Archive home page | Return to MIDRANGE.COM home page

This mailing list archive is Copyright 1997-2019 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].