How fix bug AS-162531?

[Yuriy1]

"S-162531 - On AFxx/HFxx array platforms, the NVDIMM firmware contains a configuration that will result in failing to ARM after a threshold of 5 years of total power-on time has been exceeded on a given NVDIMM. Once the threshold has been crossed, services will fail to start following a planned or unplanned reboot of an affected controller."

[Vipin_S]

Hello,
NimbleOS 5.2.1.1100 or 6.1.1.300 has been released with a fix to this issue. 
So please upgrade the Nimble OS version after going through the release notes of the 5.2.1.1100 or 6.1.1.300 because it contains all the pre-requisites for the activity
https://infosight.hpe.com/InfoSight/media/cms/active/NimbleOS_Release_Notes_5.2.1.1100.pdf
https://infosight.hpe.com/InfoSight/media/cms/active/pubs_HPE_Alletra_6000__HPE_Nimble_Storage_Array_OS_6_0_0_300_Release_Notes_6_0_0_300.pdf

[ewwhite910]

What happens if the threshold is crossed and the NVDIMM and controller servces stop?

My thought is that would create a situation where the controllers won't be able to apply the update. 
Is there a fix that does not require new NVDIMMs or a controller replacement?

[buzzsubash]

Hi,

Once the threshold is reached, controllers will fail to boot. This is a critical bug/issue.

We have a fix integrated in  latest versions of Nimble OS.

[ewwhite910]

Is there any action available for a system whose controllers both hit this threshold and now fail to boot?

I have two customers who either didn't upgrade firmware or understand the threat to their HF40 systems.

[buzzsubash]

The patching period is over, now the guideline is asking customer to update Nimble OS. But there are always exceptions including the situation you just mentioned where controllers are down or arrays in dark site etc.

We, at the support team can assist in patching the array to bring the controller back online. Please note, this can be done only on arrays that are still under valid warranty/support contract.

Please do create individual support cases referencing their serial number. Alternatively, you may directly call the support line 24/7, an L3 engineer will be readily available to assist over a zoom session.

[ewwhite910]

One is an air-gapped site and has both controllers down. That requires some intervention.

The other site had HPE support engaged, and engineers helped run the upgrade for one controller within the patching period, but did not assist with the failed controller that was already impacted. In the time since, the customer's support lapsed and HPE is unwilling to assist.

[buzzsubash]

Understood. 
Unfortunately, there is nothing much we could do without the the support contract renewed.

[ewwhite910]

That seems like a defect or recall-level fault in the Nimble solution.

[buzzsubash]

It is not.

Arrays running only a certain model of NVDIMMs are affected. And it is not a hardware fault. 

[ewwhite910]

I can help you fix this if you still need assistance. 

[Truklhats]

So let me try to understand this; HPE has sold and provisioned a product into a customer environment where there is a known timed event which will shutdown their Nimble array. This is guaranteed to happen should the customer not have a support contract and thus cannot access the firmware or OS update to remediate the known issue?This places the customers' business at risk. The only fix for a design problem is a paid contract or upgrade? Please advise. Thank you.  

[ewwhite910]

It's a bit bad.

[ewwhite910]

Or an independent engineering team who can assist...

Please reach out!