There are of course loads of ways to avoid this happening. I my case here I am going to follow a best practice recommendation for TMG and move my logging to a non system drive. The TMG BPA recommendation is the following"
"Forefront TMG is configured to write logs to the Windows system drive. If this drive is filled to capacity with logs, the server will crash and fail to start. "
There are of course hundreds of different ways that physical drive and volumes can be presented to the operating system. In my scenario the server are physical server with two physical drives. The drives are configured in the RAID controller as a single RAID 1 volume for full fault tolerance.
This presents one disk to the OS. Typically during install the whole disk is provisioned with a single partition for the OS.
The steps below will shrink the primary OS partition. This will free up space on the disk to create another partition that can then be used for a log store drive.
Shrink the System partition to create space
- Open Computer Management
- Select the Disk management node
- Right Click the C: partition and select Shrink Volume
- From the menu screen specify the amount of space to shrink
- Wait for the process to finish
You should now see the disk showing one healthy partition and another section as "unallocated"
Create the secondary drive
- Right Click the un-allocated space
- Select new simple volume
- Specify the amount of space to allocate
- Assign a drive letter
- Format with NTFS
- Finish the wizard
You should now see the disk showing two healthy partitions
Next up you want to change the service's logging location to the new drive
Move TMG Logging location
- From the TMG management console
- Select Logs & Reports
- From the Tasks pane select Configure Firewall or Web Proxy Logging
- Click the Options button
- Specify "This folder"
The logs will now be stored on your second partition. Even if this fills up it will no longer cause your server to fall over.
Correctly configuring logging on a server can reduce potential problem related to disk space. If this was not done at build time or if there are other restriction you can still follow best practice using this method.
Side note: If you are using a TMG array. Complete the disk changes before making the logging changes since the change will be applied to all hosts in the array.