patches for other systems show up

[Jeff Sadowski_2]

When submitting inventory.xml created by swainv
and checking all patches that are applicable to my system I get a few that don't seem to apply still.

It recommends the following patches still

PHKL_34881 PCI Error Handling feature; Dump A-A
PHKL_34883 Dump support for Active Active arrays
PHSS_29649 rx5670 System Firmware Rev 03.23 Patch
PHSS_32822 rx2600 System, BMC & iLO MP Firmware Patch
PHSS_33092 rx16x0 System, BMC & iLO MP Firmware Patch
PHSS_33096 rp4440/rp4410/rp3440/rp3410 iLO MP E.03.15
PHSS_33891 rp3410/rp3440 PDC45.44/BMC03.51/MPE.03.15 FW
PHSS_33893 rp4440/rp4410 PDC45.44/BMC03.52/MPE.03.15 FW
PHSS_34022 rx4640 System, BMC and MP Firmware Patch
PHSS_34343 BL60p System, BMC & iLO MP Firmware patch
PHSS_34608 OV EMANATE15.3 PA RISC Consolidated Patch 4
PHSS_34995 rx2620 System, BMC & iLO MP Firmware Patch

I have an rx2600 most of these don't even apply because of that alone. So I'm wondering why it is telling me I should apply these patches?
Is there something I need to put so that swainv will create an inventory.xml that will show itrc that I don't need the proceeding patches?

Also on the PHSS_32822 patch when trying to install this patch It says I have the same firmware as it would apply.

[Jeff Sadowski_2]

the two PHKL ones applied and are no longer shown on my list the rest are still there though

[J Goodwin]

I have also noticed this behaviour whilst attempting to patch our Disaster Recover system at a remote site. This server is called hpfiscaldr...

When I run a patch assessment after running swainv and uploading to the HP site, it always recommends installing the GOLDQPK from December 2006 even though this has already been applied to this particular server.

One possibility - our server names are similar in the first 8 chars - namely hpfiscal and hpfiscaldr. I've noticed that the system name as reported by the swainv script and the inventory.xml file contain the identifier hpfiscal, instead of hpfiscaldr. I wonder if this might be the cause of the confusion ?

Any ideas anyone ?

[A. Clay Stephenson]

If you do a man 1 uname you will find that node names (system names) are restricted to UTSLEN-1 and if you do a little digging in /usr/include/sys you will find that that value is 9 so that the maximum nodename length is actually 8. If you exceed this limit (which has been in place for decades), all sorts of compatibility issues result.

[Bob E Campbell]

The firmware patches are easily explained. Patches must patch something, and in SD hardware itself is not a thing. Looking at PHSS_29649 for example the ancestor is set to the fileset OS-Core.KERN-RUN. This means that as far as SD is concerned, this patch is meant for *everywhere*.

Scripts are used to control when the firmware is actually put into place. While this would not have been the way I would build something now, this has certain advantages such as IUX golden images including firmware for many different systems.

[J Goodwin]

A reply from the legendary A Clay Stephenson - I am honoured !

I knew of the hostname problem a while back as our previous login script gave the impression we were on the same system i.e. hpfiscal as it was performing uname -n. Hostname shows hpfiscaldr. I simply changed the script but was aware that the hostname wasn't quite right.

I did go through the inventory.xml file and changed the name from hpfiscal to hpdr in a vain attempt to get it recognised but the check routine on the website still said I needed to install the GOLD QPK from Dec2006 although it is already installed.

I will have talks to see if we can get the system hostname changed to something more sensible. Thanks for all the input.