OV4VC 11.5 + OV 8.70 + vLCM = still no reliable patching

[pirx]

I've not yet given up... but I probably should... First tests with OV4VC 11.5 and hosts in monitored mode showed that the force install bug was fix - wooohhooo.

But updates via vLCM and OV4VC still fail ~50% of them time. In below case I started update for a small cluster with just one host and it was successfull. Then I continued with the remaining 3 hosts. 8 firmware updates were needed. For the first of the three hosts the update failed in vLCM. In ILO all fw were updated successful. In OV old versions are shown and the update task is still active, so I can't do a manual refresh. I also can't stop this update task. Which means the update in now completely stuck until a timeout happens (hopefully).  I even updated all ILOs to latest version outside of vLCM/OV before.

[TVVJ]

Hello,

As the firmware have been updated as shown from the iLO's screenshot and the error indicates that HPE OneView experienced communication issue, if there are , no other tasks to be performed, then, could the firmware updates be stopped and then the appliances manually refreshed?

Regards,

[pirx]

Where should I stop the update? In OV there is no way to stop it, in vLCM is already failed.

[yesheswini_s]

Hi,

The steps to cancel a firmware update are given in this document. Please use this link.

[pirx]

I know how this works, but please read my initial post. I use monitored mode and therefor have no profile. There is no option to cancel the task. I had a session with a HPE consultant last year and he confirmed that cancel is not always available. I adressed my concern as there were servers that were in updating state for hours and you never know if there are rebooted again or not (this was not with vLCM but also in monitored state)

TheCancel taskoption is available for a task only in the following scenarios:

• When updating firmware with a server profile with the Firmware only option.

• When a firmware update operation has already begun.

[Rama2]

The Firmware files are pushed from Oneivew SPP baseline to ILO and the flashing happens using from HPE ILO repository and once the flashigh process is initiated there is no way to stop it,
If its failing on few server, we need to review the logs to diagnose the issue

[pirx]

@Rama2 The update was done in ILO hours ago and OV still thinks its ongoing. So there are 2 issues:

- OV does not get the updated versions, ILO has the correct state

- OV just keeps the update task active, even though it finished successfulls 

So communication between ILO and OV seems to be broken. I've still other cases related to OV open, never got much feedback.

[Rama2]

Hi

Kindly reboot the oneview and check the status, if still the same. Try to reset the ILO from oneview and check if that resolves the issue. 

[pirx]

@Rama2 Tried both, didn't solve the issue.

[Tinux]

I am with @pirx, the proces to update firmware from vLCM with OV4VC, whether you have profiles or not, is not smouth.

I have the problem that the Smart Array firmware for E208i is not updating at all, while OV shows it needs to be updates in the FW compliance section.
Current FW is 5.32 and needs to go to 6.52.
When trying to run remediate AGAIN after the same error @pirx expirienced, the staging proces in vLCM just hangs.
I waited 20 minutes and canceled the task ... which eventually (15 minutes later) did.

I tried updating with SPP2023.09.03 which contains 6.22 but it does not contain 5.61 nor 5.32 which are prerequisites when you're updating from lower versions ... why are these bundle's removed from the SPP !?!

So I was forced to use SPP202309.04 which contains 6.52, which can update in one step from 5.32, but it also implied that I needed to manually add the latest SmartPQI driver to the vLCM Image otherwise ... vSAN not compliant !

Conclusion: I will have to manually upload the HPE_SR_Gen10_6.52_A.fwpkg to the iLO and manually start the install.

You're better of with running SUM on a host, point it the iLO's, do the inventory and deploy.
OV and OV4VC are useless

[pirx]

And it's not getting better. Updated to OV 8.80 + OV4VC 11.6 a while ago (OV 8.90 is not working at all with vLCM).  Still only 50% (max!) updates are successful. And it seems that there are timing issues or other things that break updates. Sometimes the update was sucessful but the compliance check at the end fails. If I just rerun it manually, the server is suddeny compliant. Maybe removing a hardcoded sleep here and there and using a better logic could improve things a lot.

Update that was started in VMware vLCM was done 100% successul in ILO

All updated firmwares in ILO have the expected version

OV is aware about the updated versions

But the HSM plugin in vLCM still reports the server as non-compliant and the the update process stopped after the first host.

An imediate new vLCM compliance check then results in a compliant host. WTF!