In the last several weeks, I’ve been having a great time writing PowerShell functions and modules for new projects moving to Amazon Web Services (AWS). I’m thrilled with the inclusion of UserData as a part of provisioning an EC2 instance. Having developed my PowerShell skills, I’ve been able to leverage them in conjunction with UserData to do all sorts of things to my instances. I’m reaching into S3 for installers, expanding archive files, creating folders, bringing down custom written modules in UserData and invoking the contained functions from them there, too. I’m even setting the timezone. It’s seems so straight forward sure, but getting automation and logging wrapped around that need, is rewarding.
As a part of an automated SQL installation — yes, the vendor told me they don’t support AWS RDS — I had a new challenge. It wasn’t overly involved by any means, but it’s worthy of sharing, especially if someone hits this post in a time of need, and gets a problem solved. I’ve said it a millions times: I often write, so I have a place to put things I may forget, but truly, it’s about anyone else I can help, as well. I’ve been at that almost three years now.
Back to Microsoft SQL: It’s on an ISO. I’ve been pulling down Zip files for weeks, in various projects, with CloudFormation, and expanding them, but this was a new one. I needed to get at the files in that ISO to silently run an installation. Enter the Mount-DiskImage function from Microsoft’s Storage module. Its help synopsis says this: “Mounts a previously created disk image (virtual hard disk or ISO), making it appear as a normal disk.” The command to pull just that help information is listed below.
PS > (Get-Help -Name Mount-DiskImage).Synopsis
As I typically do, I started working with the function in order to learn how to use it. It works as described. Here’s the command I used to mount my ISO.
PS > Mount-DiskImage -ImagePath 'C:\Users\tommymaynard\Desktop\SQL2014.ISO'
The above example doesn’t produce any output by default, and I rather like it that way. After a dismount — it’s the same above command with Dismount-DiskImage instead of Mount-DiskImage — I tried it with the -PassThru parameter. This parameter returns an object with some relevant information.
PS > Mount-DiskImage -ImagePath 'C:\Users\tommymaynard\Desktop\SQL2014.ISO' -PassThru Attached : False BlockSize : 0 DevicePath : FileSize : 2606895104 ImagePath : C:\Users\tommymaynard\Desktop\SQL2014.ISO LogicalSectorSize : 2048 Number : Size : 2606895104 StorageType : 1 PSComputerName :
The first thing I noticed about this output is that it didn’t provide the drive letter used to mount the ISO. I was going to need that drive letter in PowerShell, in order to move to that location and run the installer. Even if I didn’t move to that location, I needed the drive letter to create a full path. The drive letter was vital, and this, is why we’re here today.
Update: See the below post replies where Get-Volume is used to discover the drive letter.
Although the warmup here seemed to take a bit, we’re almost done here for today. I’ll drop the code below, and we’ll do a quick, line-by-line walk through.
# Mount SQL ISO and run setup.exe. PS > $DrivesBeforeMount = (Get-PSDrive).Name PS > PS > Mount-DiskImage -ImagePath 'C:\Users\tommymaynard\Desktop\SQL2014.ISO' PS > PS > $DrivesAfterMount = (Get-PSDrive).Name PS > PS > $DriveLetterUsed = (Compare-Object -ReferenceObject $DrivesBeforeMount -DifferenceObject $DrivesAfterMount).InputObject PS > PS > Set-Location -Path "$DriveLetterUsed`:\"
Line 2: The first command in this series, stores the name property of all the drives in our current PowerShell session in a variable named $DrivesBeforeMount. That name should offer some clues.
Line 4: This line should look familiar; it mounts our SQL 2014 ISO (to a mystery drive letter).
Line 6: Here, we run the same command as in Line 2, however, we store the results in $DrivesAfterMount. Do you see what we’re up to yet?
Line 8: This command compares our two recently created Drive* variables. We want to know which drive is there now, that wasn’t when the first Get-PSDrive command was run.
Line 10: And finally, now that we know the drive letter used for our newly mounted ISO, we can move there in order to access the setup.exe file.
Okay, that’s it for tonight. Now back to working on a silent SQL install on my EC2 instance.
Would something like this work:
$driveLetter = (Mount-DiskImage “$isoPath” -PassThru | Get-Volume).DriveLetter
That’s awesome, Abhi, and really, I think I prefer it. In fact, I’m changing my code (in the project in which I’m working). I’m kind of not surprised that I overlooked Get-Volume. I didn’t see a drive letter output by Mount-DiskImage and went to that in which I had the most experience: the Get-PSDrive cmdlet. I suppose I could use the first below example, although it might not be the friendliest for someone else. I did, however, compare what I have below to what I had originally, and there’s no real time loss between the two.
Set-Location -Path “$((Mount-DiskImage -ImagePath ‘C:\Users\tommymaynad\Desktop\SQL2014.ISO’ | Get-Volume).DriveLetter):\”
If I change the above example to the below example, I could put the current location on the stack AND move into the ISO in a single command. This, making it easy to leave the ISO’s drive letter when I’m done and ready to dismount the ISO with Pop-Location. So much fun!
Push-Location -Path “$((Mount-DiskImage -ImagePath ‘C:\Users\tommymaynad\Desktop\SQL2014.ISO’ | Get-Volume).DriveLetter):\”
As always, thanks for your input!