So, here is the scenario I was faced with last Friday;
I was asked to look at a development servers (HP Proliant DL380 G6)(in other words – get it working and updated) as one of our key applications was throwing errors and the development system was required for testing. The development server had not been touched/updated in a few months and I knew there were a couple of outstanding issues on the system already. I tried to log into the system and I was presented with an authentication issue.
Update all the drivers to the latest HP Support Pack, install any missing Microsoft security patches and resize the C: drive as there was only 600MB free, which is not a lot when you are running Windows Server 2008 (not R2).
There is one array containing 6 x 146Gb disk, there are 4 logical volumes ; c: d: e: f: – C is RAID 5 and the rest are RAID 50.
There is not enough space on the C: drive to install any updates, there is no way to resize the logical C drive as all the remaining logical disks are located after the system disk.
What I did and what I experienced:
First of all, I couldn’t expand the C: drive after I had shrunk the logical drive next in the list (D:) within the array configuration utility. I then moved all the data from D: and F: onto E: and backed up E: to tape. (D and F were small in size)
I deleted each of the logical disks D, E and F and then tried to resize the C drive. To my surprise the option to expand the logical drive was not there. So… I deleted this logical drive also.
Now I have a server with 6 blank disks and no array configured. At this stage I realise that I am really going the long way about fixing the server. I start again from scratch, I add two additional disks, I create an array with 2 disks (RAID 0 – mirrored) and run a BMR (Bare metal recovery) on the C drive from the most recent backup – thankfully our development systems are included in our backup plan).
The restore was successful and the server came back but needed to be removed and re-added to the domain – its trust with the domain was broken.
The server was added back to the domain (following the steps below):
- Login to the server with a local administrator account
- Add the server to a work-group and restart the server
- remove the server computer account from the domain
- Add the server back to the domain and restart the server
Once the server was accessible again, I needed to add back in the additional disks. Then I was ready to do another restore, of the date from the 3 remaining drives.
the process does sound very long winded but I am highlighting all the areas that other may fall victim to. #1 rule: make sure you have recent backups
So, now I have a working system, on the domain, with newly resized disks and enough space on the C drive to install the necessary OS (Operating system) security updates and the latest drivers.
I ran into another issue while updating the drivers – that is that the NIC drivers installed but the devices came up with the error “This device cannot start. (Code 10)”. This caused more grief as I was working on a server in a remote location but thankfully I had iLo access. I had to download the latest firmware for the NIC and mount the folder that contained the drivers through iLo. This folder was then presented to the server and I could continue with the firmware update. This paragraph of work took me nearly a day with all the playing around with driver versions, as I ended up installing 4 different versions until I realised it was a firmware issue. Just in case you find yourself in a similar situation.
The server is finished and is fully patched with the latest security updates and drivers. There is now over 50GB of free space on the system disk.