Jump to content
RESET Forums (homeservershow.com)
Sign in to follow this  
LoneWolf

Odd disk layout after first-time WS2012E install..any advice?

Recommended Posts

LoneWolf

I'll start off with my hardware configuration, so I can provide as much detail in my setup:

 

HP ProLiant Microserver N40L, 8GB RAM

HP Remote Access Card

HP Smart Array P410 PCIe RAID controller card with 256MB cache (the drive backplane is attached to this rather than the mainboard mini-SAS port)

(3) Samsung HD204UI 2TB drives, (1) WD Red 2TB drive, in a single RAID-5 array (I used the WD Red because the HD204UI drives are no longer available, and I needed a fourth. Capacity does match, the WD is the first drive in the array)

HL SATA DVD-ROM (connected to the single SATA port on the system board)

 

I also have an external 2TB Hitachi HDD attached via eSATA, but this should be irrelevant, as the issue appeared prior to hooking it up.

 

I set up Server 2012 Essentials from a bootable USB key, with the intention of having a modest C: drive (150GB in this case), and the rest of the array in a single large volume (I hadn't decided on NTFS or ReFS yet). Creation of the RAID-5 array in the controller BIOS went off with no issues or errors; in fact it defaulted to the exact setup I wanted. I did the advanced install of server and created my C: partition, MS of course, created the System Reserved partition as well. I decided not to create additional volumes until Server was fully installed.

 

The install went okay, no serious issues there, but when I go to Computer Management and look at Storage, Windows set my volumes up like this:

 

WS2012E_diskmap.jpg

 

 

I have no idea why Server decided to put in two volumes after the C: drive. I can't delete either one. While I can create a simple volume on the first area of space, I can't even do that with the 3TB option.

 

My understanding with WS 2012 was that I could create huge drives if I wished, and this just makes no sense to me. It might make more if the volumes were broken into roughly 2TB chunks; I'd think it had to do with earlier file system limitations, but that doesn't appear to be an issue. I also don't know why I can't create a partition in the second volume.

 

Does anyone have any ideas or advice? I'd greatly appreciate it.

Edited by LoneWolf

Share this post


Link to post
Share on other sites
Boris

I believe you'd have to convert your disk to GPT to support over 2TB, but then you can't boot from GPT? Sorry, you were looking for ideas, that's mine.

Share this post


Link to post
Share on other sites
LoneWolf

Boris,

 

Your answer was a good one, while you were posting it, I found some more information along those lines based on error messages I got when trying to re-set up Windows Server:

 

http://mytechweblog....01_archive.html

 

I'm used to working with Dell PERC RAID controllers; with those, you can set up multiple containers or logical drives in an array right from the RAID controller BIOS. In addition to the GPT issue, I was confused because the Smart Array BIOS doesn't let you do this; you have to use HP's offline version of their Array Configuration Utility, which boots from an ISOLinux CD.

 

Doing this, I was able to create a 160GB logical drive and a 5.3TB logical drive. I converted the 5.3TB drive to GPT from the Server 2012 command prompt during the install with DISKPART, and have restarted setup. Once complete, I can hopefully convert the second logical drive to GPT and create it as I wish. I'll update with what I find.

Edited by LoneWolf

Share this post


Link to post
Share on other sites
LoneWolf

The above solved the issue. Two logical drives on a single RAID-5, the boot drive as an MBR volume, and the secondary logical drive as GPT. From what I have since read from Microsoft, it appears I would have had to do it this way no matter what, unless I made multiple logical drives each 2TB or smaller in size accross whatever arrays I created.

 

My only surprise is that Server 2012 (which almost certainly means the same for Windows 8) doesn't make this process more straightforward during installation; it isn't like single drives larger than 2TB weren't available prior to its release. Instead, there's a KB article or two that explain you have to manually convert logical drives to GPT with DISKPART, rather than the ability to do so during the GUI portion of the advanced partitioning setup while installing. I also had to do the conversion once Windows was installed due to not being able to back up from the command prompt in Repair mode to the installation screen (restarting after running DISKPART in Repair mode somehow reversed my changes each time). An IT pro or determined enthusiast could figure this out with a little work, but the majority of people would be lost.

 

Hopefully this thread serves as helpful to someone else doing the same thing down the line. Thanks, everyone.

Share this post


Link to post
Share on other sites
ikon

Good to hear you got it working. Now, don't be like your nickname and abandon us :D

Share this post


Link to post
Share on other sites
Joe_Miner

Boris,

 

Your answer was a good one, while you were posting it, I found some more information along those lines based on error messages I got when trying to re-set up Windows Server:

 

http://mytechweblog....01_archive.html

 

I'm used to working with Dell PERC RAID controllers; with those, you can set up multiple containers or logical drives in an array right from the RAID controller BIOS. In addition to the GPT issue, I was confused because the Smart Array BIOS doesn't let you do this; you have to use HP's offline version of their Array Configuration Utility, which boots from an ISOLinux CD.

 

Doing this, I was able to create a 160GB logical drive and a 5.3TB logical drive. I converted the 5.3TB drive to GPT from the Server 2012 command prompt during the install with DISKPART, and have restarted setup. Once complete, I can hopefully convert the second logical drive to GPT and create it as I wish. I'll update with what I find.

The above solved the issue. Two logical drives on a single RAID-5, the boot drive as an MBR volume, and the secondary logical drive as GPT. From what I have since read from Microsoft, it appears I would have had to do it this way no matter what, unless I made multiple logical drives each 2TB or smaller in size accross whatever arrays I created.

 

My only surprise is that Server 2012 (which almost certainly means the same for Windows 8) doesn't make this process more straightforward during installation; it isn't like single drives larger than 2TB weren't available prior to its release. Instead, there's a KB article or two that explain you have to manually convert logical drives to GPT with DISKPART, rather than the ability to do so during the GUI portion of the advanced partitioning setup while installing. I also had to do the conversion once Windows was installed due to not being able to back up from the command prompt in Repair mode to the installation screen (restarting after running DISKPART in Repair mode somehow reversed my changes each time). An IT pro or determined enthusiast could figure this out with a little work, but the majority of people would be lost.

 

Hopefully this thread serves as helpful to someone else doing the same thing down the line. Thanks, everyone.

 

Thanks for the post and especially the links -- pretty cool. What kind of performance are you seeing from the logical drives?

Share this post


Link to post
Share on other sites
LoneWolf

Performance is fine on the first logical drive (160GB). Parity is still being built on the 5.3TB unit, it's at about 50% done. I'll be in better shape to analyze performance once that's complete.

 

Note that my drives are 5400rpm (temperature and NAS reliability were my priority over performance), and I'm using a 256MB cache module with no battery backup on my RAID controller. If I went to 7200rpm, added the 512MB cache module, and the battery backup (which would make write-back cache a no-brainer, though I could currently override the controller settings and force it) I could also up the performance some.

 

As the battery backup for the controller is kind of expensive, and I have a UPS, I'm thinking about whether I might just force the write-back cache on in the controller and let my UPS handle power events.

Share this post


Link to post
Share on other sites
ikon

I think you should be safe with the UPS, as long as the UPS properly signals the server when a power outage occurs and the server shuts down before it crashes. I always test this with my systems by actually pulling the plug.

Share this post


Link to post
Share on other sites
LoneWolf

The UPS (an APC BackUPS ES 750) has USB connectivity to the server; I've set up PowerChute Personal (which does appear to work with Server 2012 Essentials, as long as the .NET 3.5 feature is installed, as .NET 2 is a requirement).

 

I'll also do the testing you mentioned.

Share this post


Link to post
Share on other sites
ikon

There's nothing quite like having the peace of mind of knowing your server will shut down gracefully in the event of a power failure.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  



×
×
  • Create New...