A key part of most personal data backup strategies involves backing up data to an external USB drive but I don’t want to leave it constantly connected. In this post I cover how to backup to an external drive using a scheduled automated process but only if the external drive is connected at the time.
I don’t believe in leaving external backup USB drive always connected to my system (PC or Server) to avoid the data being corrupted or deleted. Also if the backup drive is for off-site storage then its not possible to be always connected anyway. Using the steps below I can perform a full data backup by simply physically connecting the drive (connecting a USB drive or docking the SATA drive into a USB dock for example), and then leaving it overnight. In the morning I can safely physically disconnect it and store it without even having to log onto the machine.
The basic flow:
1) Assuming that the machine is always on, which my server is, a scheduled task runs every night and executes a backup script (regardless of the backup drive being connected or not). If your machine is not always on then vary this by setting the scheduled task at a time when the machine is usually on.
2) The script checks for the existence of a specific folder on the connected drive (e.g: U:\backup\). If the drive isn’t connected then the folder path won’t exist and the script just exits happily. However, if the drive has been connected then the folder path will exist and the backup script will continue and copy over the data.
3) After a successful backup the script safely disconnects the USB drive. This step is technically optional as Windows supports pulling a USB drive out without doing a soft eject but its highly recommended to tell Windows first to avoid data corruption.
4) Optionally you can also output a backup log somewhere to enable you to check the logs periodically without having to reconnect the USB drive to verify the job worked.
More detailed steps:
Firstly connect the external USB drive and make a note of the drive letter it uses. Changing the drive letter to something memorable might help (B for backup, U for USB, O for Offsite etc). We’ll use U for this example. Next we need to write a DOS command script, a simple program or Powershell script to perform the backup of the data using the backup tool of your choice. I use Robocopy to copy the files via a Powershell script and below is a simplified version of my script.
#==========================================================================================================
# Checks for presence of offsite backup USB drive, and backs up relevant data to drive if present,
# exits gracefully if not present. Run script everynight and it only backs up data when offsite
# USB external drive is turned on.
#==========================================================================================================
clear-host
# set file paths and log file names
$timestamp = Get-Date -format yyyyMMdd_HHmmss
$LogBasePath="D:\Logs\OffSiteUSBBackup"
$LogFile="$LogBasePath\USBBkUp_$timestamp.txt"
$USBDriveLetter="U"
$USBDriveBackupPath="U:\Backup"
# set error action preference so errors don't stop and the trycatch kicks in to handle gracefully
$erroractionpreference = "Continue"
try
{
# Check USB drive is on by verfiying the path
if(Test-Path $USBDriveBackupPath)
{
# now copy the data folders to backup drive
invoke-expression "Robocopy C:\Docs $USBDriveBackupPath\Docs /MIR /LOG:$LogFile /NP"
invoke-expression "Robocopy C:\Stuff $USBDriveBackupPath\Stuff /MIR /LOG+:$LogFile /NP"
# Copy the log file too
invoke-expression "Robocopy $LogBasePath $USBDriveBackupPath\Logs /MIR /NP"
# Sleep for 60 to ensure all transactions complete, then disconnect USB drive
Start-Sleep -Seconds 60
Invoke-Expression "c:\DevCon\USB_Disk_Eject /removeletter $USBDriveLetter"
}
}
catch
{
# Catch the error, log it somewhere, but make sure you still eject the drive using below
Start-Sleep -Seconds 60
Invoke-Expression "c:\DevCon\USB_Disk_Eject /removeletter $USBDriveLetter"
}
It is key to include in the script a check for the existence of a specific folder on the drive letter belonging to the external drive (U in our example). Only if its present do we continue with the backup.
I make sure that Robocopy logs the output to a file and that file is on the server and copied to the USB drive (as a record of the last backup date etc). I also report the running of the PowerShell script to the Eventlog for reporting purposes but this is outside the scope of this post.
It’s safer to tell Windows that you’re gonna pull the drive out and so I call USB_Disk_Eject from within my script, passing in the drive letter. I then wait 30 seconds to ensure the drive has had sufficient time to disconnect before I exit the script. There are a few tools available for ejecting USB drives such as Microsoft’s Device Console (DevCon.exe) but I use USB Disk Ejector (https://github.com/bgbennyboy/USB-Disk-Ejector).
Now set up a Scheduled Task in Windows to run the script every night at a set time. As the script is scheduled to run every night all I have to do if I want to perform a back-up is connect my backup drive and leave it until the morning. The script will run overnight, find the drive, backup and disconnect. In the morning I can just physically disconnect the drive safely without having to log onto the machine. Periodically I’ll check the backup logs and the backup drive to make sure all is well and to check remaining drive space etc.
UPDATE: A working Powershell script can be found on my GitHub site.
Do you like this approach? Got a better idea? Let me know via the comments.
RichHewlett.com 