Oneview for vCenter: from Basline to VLCM

[erki72]

Hi,

After we are on vCenter 7 I would like to switch to VLCM.
I have Oneview 8.0. Oneview for vCenter 11.1 and vCenter 7.0.3

Initial situation: all servers are in Oneview and have a server profile there (which was created from a server profile template).  We use this to control the BIOS settings and the SPP.

Now I would like to distribute the SPP and Vendor Addon via the vCenter vLCM. What do I have to do with my profiles in Oneview?

[support_s]

System recommended content:

1. HPE OneView for VMware vCenter 11.1.0 Release Notes

2. HPE OneView for VMware vCenter 11.0 Installation Guide | Checklist for HPE OneView for VMware vCenter Installation

Please click on "Thumbs Up/Kudo" icon to give a "Kudo".

Thank you for being a HPE valuable community member.

[erki72]

Hi @support_s 

Thanks for the links.
The installation is done and the plugins are visible in vCenter.

My concern is how to get rid of the server profile in Oneview after creating the image.
I was able to create the image in vLCM.
Can I now easily use the image without having to do anything with the Server Profile Templates in Oneview beforehand, although the SPP still comes from there?

[T_1_6]

In my experience you can do what you like.

If you do nothing with the server profiles, VLCM will update things, (whether it works flawlessly is the subject of debate), and all your server profiles will warn that the firmware is non compliant.

If you edit the server profile/template before hand, and follow HPE instructions/recommendation, then the firmware gets distributed out to the servers in preparation for the update during vLCM because of the settings in Oneview.

To be honest I have not found it actually makes any difference either way, as the vLCM stuff has been super flaky for us, requiring many attempts at updating the cluster, before all the servers eventually stop erroring, and reach the desired update state. Likely something to do with ILO firmware, or whatever, I dont know, and I am not the only person either.

[erki72]

Hi @T_1_6 
Thank you very much for the feedback which is very exciting to read.

Oops, and I was hoping it would get better and more importantly easier.
Are you still doing it this way or have you switched back to the "old" way?

I will test it with a cluster and set it in the profile template beforehand so that the firmware does not have to "exact match",

[erki72]

Good day @T_1_6 

Yesterday I changed the profile templates so that the baseline is no longer checked. 
Then I wanted to create an image baseline in the vCenter. Well, I'm not exactly convinced by this either.
Whenever I select the vendor addon, the baseline cannot be created.
It looks way too static to me and I don't think I'm flexible enough either, so I'm wondering how to continue

For me, the big advantage would have been that the hosts only reboot once to update ESXi and HPE drivers / firmware and I can do the whole cluster automatically.
Now I'll probably go back to Baslines, firmware from Oneview and which means I'll do it again per host with multiple restarts
Or is there something I'm forgetting?

CC: @support_s 

[T_1_6]

Dunno man, the vLCM stuff and the cluster image is pretty slick.

When you are done editing the base image you can validate and it will tell you if its fine.

You need to register the SPP with vLCM too remember, via the plug in.

Its been a mixed bag for me, Ive used it for 2 years now, and have never had a single sucessful reboot upgrade. Always multiple, always something fails, and I have to keep remediating the cluster over and over until eventually all hosts succeeed.

Aside from that its been quite good.

[erki72]

That's as far as I'm actually.
Registered the SPP in the vLCM and see that too.
As soon as I select the vendor addon, I cannot create the image baseline.

Anyway, if i read your lines, it's not really different than i have now with the multiple reboots per host.
I find that very interesting because we have a lot of hosts that are configured exactly the same, and yet something is always different

[T_1_6]

Never had a problem selecting vendor add-on myself, thats a Vmware problem, not an HPE one.

Only issues being when the custom add on is not allowed initially because of a downgrade message with the applied SPP having a newer driver than the custom add on.

Easily resolved though by customising the individual components.

[erki72]

Oh that's exciting now. That's probably what my current step is all about.
The message: "The operation cannot continue due to downgrade of the following Components: MRVL-QLogic-FC".

Then I would have to bring the current driver to the addon depot, do you have a link for me?

[T_1_6]

You can arse about manually finding the exact driver to override, OR, you can save your sanity and simply select one back from latest HPE custom add-on, and it will likely work fine.

[T_1_6]

In production I run:

ESXi Version 7.0 U3i - 20842708

Vendor Addon  HPE Customization for HPE Servers 703.0.0.10.10.0-12

Firmware and Drivers Addon  Gen10 Service Pack for ProLiant SPP-2022.09.01.00.703

[pirx]

Maybe this helps as I was running into the same problem.

vLCM and Image based update (error: Downgrades of ... - VMware Technology Network VMTN

For me vLCM + OV + SPP does not work due to a different error. But ILO Amplifier as HSM works....

ov4vc 11.1 + SPP 2022.09.01 + vLCM = two manifests... - Hewlett Packard Enterprise Community (hpe.com)

[erki72]

Thanks @pirx 

Every hint helps me

This is something i could try.
Can you just say do I need to disconnect OV4vCenter first or can I just connect to the vCenter in Amplifier?

[pirx]

If you don't get an error like I get during patch download / repo sync in vCenter, you don't have to switch to ILO Amplifier. Maybe this is an isolated issue in my setup.

Regarding the downgrade issues look at my first link, if you go through the list of error'ed packages until all are covered, the image should be fine.

[erki72]

Okay, then I'll leave it as it is. Because it really seems to be just a driver..

How do I find out which one it is or how do I get the driver from Addon Depot?
Use esxcli to see what's installed?

[pirx]

Did you check my fist link? Search in Component -> Add Components -> Independent Components and Vendor Addon Components for the package/driver that has errors. You can use the filter in component name column, but the name there is not always identical with the name in the error. But you could start with searching for mrvl or qlogic and then see if a package with the right version number from the error message is available and choose this.

[erki72]

Yes i saw the link thanks
Have to go through some, haven't found it yet. I'll take my time.

[ChrisLynch]

Folks, the iLO Amp HSM provider is deprecated and will not be supported.  The last supported vSphere version was 7.0U2 with vSAN, 7.0U3 without vSAN and will not support vSphere 8 or newer.  Gen10 and newer vLCM customers need to transition to using HPE OneView.

As for some of the original questions about using server profiles and templates, yes you need to with HPE OneView.  We will have an updated OV4VC release that will support HPE OneView Standard customers with at least HPE OneView 8.0 or newer.  It will then perform server hardware firmware updates that was introduced in HPE OneView 7.2 (offline SPP update only), but HPE OneView 8.0 added vLCM online firmware deployment support.

[pirx]

That means that in future we don't need a license for OV to be able to update firmwares? With or without vLCM. But currently if you do not have a OV license the only option is ILO Amplifier as HSM, right?

[erki72]

Good morning @ChrisLynch 

I heard about iLO Amp HSM.

Now we have vCenter 7.0.3, OV 8 and OV4VC 11.1
Nevertheless, it seems to me like a botch to make the switch to the vLCM.

As usual, I will change the SSP in Oneview in the profile template, patch the servers in vCenter with baselines and after the reboot the SPP will be applied from Oneview because of the server profile.

Yesterday I manually added the latest SPP to a cluster so that I might be able to coexist in vLCM. Unfortunately still the same message "The operation cannot continue due to downgrade of the following Components: MRVL-QLogic-FC."

[T_1_6]

This has already been convered and you have the answer.

Either switch out the driver version using the manual override, or, go back a version in the HPE Vendor add-on.

Not sure why this is still an issue.

[erki72]

Yes they wrote and mentioned that.
However, I can't find the driver in the Vendor Addon and I don't know how to do a manual override either, because I don't know where and what it is.

I really don't know why the problem exists. Can now also see that the firmware on the "Marvell FastLinQ 41000 Series - 2P 10GbE SFP+ QL41132HLCU-HC" adapter has been updated.

[pirx]

It is not allways obvious as the name in the error and the package name in AddOns are not same. If you search fpr MRVL you will not find anyting. But if you either seatch for qlogic in Component Name or for Marvell in Venodor, you will probalby find the package. This is not straigth forward and it's a pitty that it has to be so complicated, but that the new hot....  I'm just wondering why your error does not show a package version, when I've such an error it tells me the versions so that I then can choose the right one from the override.

[erki72]

It's very tricky.

Compare both addons:

the older addon is valid and could probably be installed.
If I then compare the HBA driver (I assume it's because of him) I don't have the version at all