HPE Composer2 Stuck on date (1st January 2017) -> Impossible to reimage

[ConstantinH]

Hey team, 

facing quite a strange problem. We are trying to reimage a Composer 2 with version 6.60.04.

Everthing goes well until one point where we get the message "Could Not Validate the HP Signature Because the HP Public Key Could Not Be Imported". We are already aware about the advisory : https://support.hpe.com/hpesc/public/docDisplay?docId=emr_na-a00137087en_us
Our problem is that the date inside of the composer 2 is stuck on the 1st January 2017. For a reason we ignore, there is also no more OneView on the Composer 2. When we but, we arrive on the "grub" command line. 

We gave the ilo an IP but no way to change the date from there. 

Do you have any clue what we can do ? 

Tomorrow we will open a ticket with the remote support

[DanCernese]

If your OneView no longer boots and you cannot perform the workaround in the advisory you found; then you must open a support ticket.   I have not heard of a "stuck" date that wouldn't allow it to be changed; that's new to me.  So if you cannot boot any longer that will mean our support won't even be able to login to the maintenance console service command line to fix the date.  It may require some new parts.

Nothing easy can be done if both the system board (where the reimage is initiated) and the ssd/disk have problems at the same time.  One or the other could be fixed if you have spares to swap or another Composer to work with temporarily rebuilding from those parts.  If you were to receive both a new system board and a new disk-- you cannot create a working composer from them because one or the other must be working to start (our composer system board spares pool still has the date issue in the advisory at the time of this writing).  Our services would need to pre-prepare a working disk in order to fix the board and vise-versa.

[mickeytheseal]

Hello!
I have an good Composer2 with active Appliance in appliance bay 1 with version 6.10.00
Oneview in appliance bay 2 was damaged, so I decided to reimage it.

I'm trying to re-image Composer2 with version 6.10.00 instrallation file.
I see an error:

error: no such device: EMBEDDED330.
Re-image of appliance from USB - this process can take up to 10 minutes.
Locating firmware image on USB: USB
Current date and time is: Sun Jan 1 00:01:06 UTC 2017
Could not validate the HPE signature because the HPE Public Key could not be imported or is not valid.
Re-image of appliance encountered errors and could not complete.

I tried changing the date on the active appliance to 2023 but it didn't help.

Maybe you can tell me how to fix this situation?

[DanCernese]

Later versions of the reimage script (embedded on the Composer system board; so not easily modified) should allow you to set the date-- but I'm afraid I don't know how to recover when it won't let you set it -- so you'll need to open a support case and explain the actions in the published advisory are not applicable.

You can't set the date on the active composer to fix  one that is not part of the composer ha cluster.  However, if you can temporarily shut down the active one and fully boot the one you want to reimage, then you can set it's date that way.   However, if the target reimage composer is so damaged that it will not boot fully then you must engage support.

So if you or the original post would share your support case# someone can at least watch it to see if this is a new issue or just different symptoms of the same issue (but it still requires support to recover).

Just for background, HPE had address this problem in both OneView updates and in our factory process.. ...but the old embedded reimage script does remain on some old Composer2s.

[mickeytheseal]

Thanks for your reply!
So, this reimage script is built into the Composer2 system board hardware?
If HPE has fixed this issue at this time, then maybe there is a newer revision of the system board? My board spare number is 879540-001.
I might try to find another system board where the reimage script problem is fixed.

On the other hand, what do you think about this method:
1. stop the active composer
2. Remove its sdd drive, connect it to the PC and clone it to the faulty one. For example, using linux dd command.
3. Install active composer
4. Install standby composer with cloned ssd

Will this work? Can such a cloned image of clone active appliance, become a standby appliance?

[DanCernese]

While your idea is sound, HPE's spares inventory is not yet reworked.  It is highly unlikely that you would receive a corrected system board in the near future because we have quite a large worldwide inventory to address.  It would be much easier to open a support case.

Also:  there is never a reason to go to such lengths to clone a disk just to become a secondary composer:  a secondary composer joins the cluster in a default configuration and syncs with the active.  So you are always better off reimaging that secondary (yes, I know reimage is your issue), but you can't build a composer from parts so you must resolve your reimage problem with the support team.

[DanCernese]

Definitely do NOT clone the drive-- it will change the Composer's serial number to match the original the disk came from and then things just go downhill from there (because the disk thinks you changed the system board, which is the supported process.  not the other way around).

[mickeytheseal]

@DanCernese 

Hello DanCernese!
In the end, I did the following:

1. I took the image "Synergy_Composer_2_6.10.00_INSTALL"
2. I unpacked it and realized that the file inside is a Linux DD image
3. I extracted the SSD from Composer, inserted it into my PC.
Then, I used Linux DD to write this image to SSD.
4. Returned the SSD to Composer, returned Composer to Synergy Frame.

HA cluster is working! Synchronization was successful.
I am attaching a screenshot.

But there is a problem - a message appeared in the activity:
The appliance encryption key is stored on the appliance hard drive. The credentials and private information it encrypts could be compromised if the hard drive is stolen.

Please tell me what I should do next?

image_1-HA-cluster

image_2-encription