Host Profile works on all but 2 blades... CP_NO_VALID_PARTITION_FOUND?!

[lennyburns]

I have successfully installed 14 of 16 blades successfully using Auto Deploy and a VERY PLAIN / Vanilla host profile.

The sole purpose is to get ESXi5 on each blade and then swing the installation to a STATEFUL cached installation (local installation).

However, two blades refuse to complete the profile application without error.

They complain about " CP_NO_VALID_PARTITION_FOUND" in the syslog.log file and even if I manually place a VMFS5 volume, it still refuses to complete without error.

This leaves these two hosts in Maint. mode and they fail to come online.

Note that my expectation is that these servers can start with a completely ERASED set of drives, where only the RAID controller has made a 0+1 disk of the two on board drives.

It SHOULD install just fine on that, but NOT on these two servers.

Attached is a the Syslog....

================================

ANY help is appreciated.

[lennyburns]

Here is better information...

I attached two logs.

One is a a freshly installed blade called ESXi05 that works perfectly - not one of the two "bad" servers.

The other is a freshly installed blade called ESXi13 (lucky number 13 is a bad one - imagine that)

They are EXACTLY identical in EVERY way i can tell, same firmware, same hardware and in the same chassis.

Both identical.

server 05 installed and joined perfectly.

server 13 went nuts.

The logs should be identical, but aren't!

WHY?!

I cant deduce (yet) what is happening to server 13!

[lennyburns]

Its baffling.

A in-depth examination of the servers using the advanced view of the HP Diagnostics report shows no material difference that wouldn't be expected (serial numbers, etc)

Attached for the morbidly curious.  File compare tools show only normal differences.

Why are these servers installing differently?

It appears that the 13 server cant make partitions or carve the drives right.

[lennyburns]

PROGRESS!!!!

I swapped the SATA drives with the bad server (13) and a perfectly working server (05).

The problem MOVED to the working server!

What in the WORLD??!

Since I have two misbehaving servers, I'm going to build two working servers, then break the 0+1 mirrors on them by removing 1 drive from each.

This will give me two drives from a mirror set of a working server.

I will remove all drives from the bad servers (four in total) and replace them with just ONE of the working server drives (being sure to place them in the same position they came from on the working servers.

Leaving the working servers that donated drives powered OFF, I will boot the bad servers on ONE drive, where each drive was from a working servers' mirror set.

Once booted, i will insert bad drives in each server and this SHOULD force the mirror to overwrite the bad drives and clobber whatever is on them that makes things go bad during installation.

I HOPE!

Once all is working, I will force them to PXE boot and pick up their own installations and overwrite the install that came with the donor drives.

I will then insert the remaining bad drives into the booted working servers and fore them to overwrite THOSE "troublesome" drives during a mirror rebuild.

this SHOULD eradicate the contents all of four of the problematic drives.

I have NO why i have to do this, because I have erased these drives several times using two different Array configuration utilities!

Results to be posted next.

[lennyburns · Accepted solution]

Overwriting the offending drives by making them rebuild targets of working servers eradicated the "issue" these drives were having.

No more problems exist, however i STILL don't know what the root cause was.

So, issue resolved - but a mystery remains.