DL380G5 Boot delay and hardware monitor issue

[lmaximo]

Hi to all and sorry for my english,

Our server:

3 x DL 380 G5 (2 x E5440, No HDD, 4 NIC, 24 GB Ram) with Esx 5.5 boot from USB key and iSCSI target as storage.

Our step:

Server 1: Update with firmware DVD 2014.02.0 - All OK and server work fine

Server 2: Update as for Server 1.

All update OK but server does not restart automatically at the end of update.

Black screen.

Nothing change if power off with button.

Only after remove both power cord, it start but with a strange issue.

During post message, normally it need 5 or 6 seconds between ROM information text and ILO information text.

Now it need about 1 minute!

After this delay, server boot and work without any issue.

Opening vSphere Client, the second strange issue: Hardware monitoring not work. It report "hardware monitoring service on this host is not responding or not available"

Excluding these two strange issue, all work fine.

Server 3: All happened exactly as for server 2.

Nothing change updating with firmware DVD 2014.02.0 (B) and updating Esxi to 5.5.0 U1

As already say, i follow same steps on all 3 server.

Therefore, i think that during delay in boot, "something" not reply fine to check and this "something" is what is wanted by vSphere hardware check.

What i can check to find the issue?

Thanks in advanced for any help

Kind regards.

Massimo

[waaronb · Accepted solution]

There was an issue with the Broadcom firmware update... it was flashing the wrong thing on certain Broadcom multifunction network cards.

On a Windows machine it was causing a problem because the NIC was no longer detected correctly. On a Linux or ESX machine maybe it would just pick up the new NIC information and continue? I don't really know, but maybe that's related to the other error message you saw?

Also, make sure you update the firmware on the ILO to the latest version which fixes the Heartbleed scanning vulnerability.

[lmaximo]

Hi waaronb and thanks for your reply.

Great suggestion!

After read your post, i have check network adapters list on two server and not find 2 Broadcom NIC that we use for test only and therefore not monitored!!

Can you help me about next step to fix this?

I have already update to Esxi 5.5.0 U1 HP custom image, but nothing change.

Thanks.

Massimo

[waaronb]

You might try running the server diagnostics at boot to see if anything is showing an error. When a server takes longer to POST than usual, it could be a sign that something's wrong and one of the components (the array controller, lights out, any add-on cards) is having a problem.

Check the integrated management log to see if it logged any errors.

[lmaximo]

Hi waaronb,

after your suggestion, i find advisory c04258304 and after read it, i open support ticket in HP.

Result: NIC (integrated in mainboard, in our server) must be replaced. There is no way to recover it.

After this "good" news and waiting replacement, we have disabled NIC from Server BIOS.

Now, start time is back to normal and Esxi hardware check is back to work.

I hope that this information can help other users.

Thanks another time for your help.

Kind regards.

Massimo