Server Partitioning

Hi all,
For the last few years since i've had a myriad of clients with preconfigured server drive partitioning issues and their boot partitions filling up (updates, WSUS, etc). My answer has been to configure all new servers with a single partition with all of the space. Can anyone give me good solid evidence or reasons why I shouldn't continue to use this method?
ElementITAsked:
Who is Participating?
 
rindiCommented:
"winsxs fills up consistently"

Windowsupdates, among other things, are linked to that folder, and it is what uses up most space. You can delete old updates using the Disk Cleanup tool (you need to add the Desktop Experience or Themes feature of your server OS for the Cleanup Tool to become available). On windows 2008 r2 servers you may additionally need the following KB to get the windows-updates into the cleanup wizard:

https://support.microsoft.com/en-us/kb/2852386

In my opinion it is much better to separate your OS from your Data either via partitioning or better, by setting up two RAID volumes. Also with today's large disk sizes RAID 5 has long been outgrown and you should never use it, as it is more risky than other RAID setups, like RAID 1, 10 or 6.

Advantages of using separate RAID volumes over partitioning is that you can use a conventional MBR disk for booting the OS, and a large GPT disk for the data, without having to bother about the unnecessary and complicated setup of UEFI BIOS and OS. Besides that, in today's world you should be thinking "Virtualization", and Virtual Servers don't use partitions, but rather multiple "Virtual Disks". This makes it very easy to add more space to a virtual disk in the future if necessary. Although you don't have that advantage on a non-virtualized environment, you might as well start getting used to that and use the same concept.

Another main advantage of separating the OS from the data in either partitions or disks is that you can backup or restore your systems more easily. You could just restore the data partition to another server and be ready again to work should you need to bring the original server down.
0
 
arnoldCommented:
Without knowing what the setup is, I.e. Single drive, multiple drive single logical volume presented by RAID controller.
The issue is not with partitioning, but erroneous choices when selecting where certain applications should store there data files.
In a case with wsus, wsusutil can be used to relocate the contents from the C: drive letter where they were erroneously placed Can be relocated to d:\ or e:\  as appropriate similarly placing other shares, etc. on the other partition and reclassifying.

The partitioning is supposed to isolate the boot drive from getting filled up by user/company/application data including swap space...... Impacting/causing performance issues.
In a multi-partition setup when a partition starts running out of space based on applications and the config, there are ... Fewer things to check, when the boot drive is starting to run out of space, the pressure is more pronounced as time goes on, performance and user compaints arise.
0
 
Lee W, MVPTechnology and Business Process AdvisorCommented:
Generally agree with arnold, but in my own words:

Well, first, WHY do you use preconfigured drives?  I never use them.  

Second, WHY are installing services and apps to the C: drive?  Of course they will fill up without proper management and configuration.

Single drives make sense for RDS servers but RARELY for any other server config.  Partitioning, when done properly, will help ensure no one service brings down your server.  It can also be beneficial for imaging, backups, migrations, and even a measure of protection from viruses (if they are poorly written and don't hunt for other drive letters).
0
Making Bulk Changes to Active Directory

Watch this video to see how easy it is to make mass changes to Active Directory from an external text file without using complicated scripts.

 
ElementITAuthor Commented:
So, configuration is generally RAID 1 or 5 presented by the RAID controller. You're right, when we come across WSUS on the C partition we move it but, winsxs fills up consistently and is not recommended to be moved as well as other system files that usually in total surpass the 60, 80 100GB recommended partitions. I just find it much more settling to know a space issue on a particular partition will be one less thing we have to be concerned about.

We never use preconfigured drives, we encounter them at clients that we take over network management. Many of these clients are single server environments, so many times the server is RDS, File server, application etc.
0
 
arnoldCommented:
The evolution of space/partition size grows as the version of OSs as well as the prevailing drive capacities.
While not perfectly, they seem to track close enough.
The difficulty to answer the question you posed is there is no single valid or encompassing that could address a significant portion of  possible implementation.

As network/speed performance as well as lower cost structures for SAN, there might be more deployment using diskless or minimal storage per server to retain some local sata while tge boot and other data is allocated from the SAN.
0
 
Lee W, MVPTechnology and Business Process AdvisorCommented:
If you never upgrade a os (which you shouldn't except in RARE circumstances) Then you are able to properly partition from the start.  My standard partition size is 120 GB for C:.  WinSXS can grow but especially in Win8/2012 improvements have been made.  With an appropriately sized and drive to begin with and best practices, you're fine.  And indeed, I think better off partitioning.
0
 
ElementITAuthor Commented:
Thank you all for your comments. None are totally compelling reasons to change my current method but I will certainly test in our internal environment.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.