i have a raid 1 array on my server, and i was surprised to see no
easy solution to setting up an email alert if one of the disks should
fail. luckily it has never happened but apparently windows
doesn't even pop up a task-bar alert if it happens.
i tried the Windows performance monitoring and alerts, and while it
can tell you the average read/writes per second, it can't tell you how
many disks are online. so i looked into WMI which has an API for
hard disks, but i encountered a problem with the API not returning the
information it is supposed to for disk status.
then i discovered DiskPart.exe, a powerful disk management utility
bundled with windows server 2003, i think you can get it for server
2000 too with an admin pack or something. you can run
DiskPart.exe from the command prompt, type: "Select Disk 0" then hit
return, next type "Detail Disk" and hit return. you should see a
list of all the volumes on the disk, along with the disk status
"Healthy". In the case of a raid array, i found it more useful to run the following commands:
select disk 0select volume 0detail volume
this outputs all the disks that make up the volume (Disk 0 and 1, in my case), and their status. Fortunately,
you can pipe the output of DiskPart to a text file, and you can also
tell DiskPart to run a list of commands from a script, so the whole
thing can be automated as a scheduled task in Windows. I also
wrote a simple c# console app to send the contents of the output file
in an email.
The .cmd file i have scheduled contains the following lines:
@diskpart /s c:\Scripts\disk_part_commands.txt > c:\scripts\disk_status.txt@SendMail email@example.com firstname.lastname@example.org "Server Disk Status" c:\scripts\disk_status.txt
The SendMail program takes in the following
parameters: from address, to address, email subject, text file
path. it is hard coded to use the localhost mail server which you
can change if you want in the source code (.Net C#).
Download the send mail console app: SendMail.exe (16 KB)
Download the send mail source code: SendMail.txt (1.26 KB)
if i was really serious about it, i would parse the text file and
read in the status directly, and only send the email if the status is
something other than 'Healthy', but i think it's nice to get an email
once a week anyway from the server, reporting it's disk status and
letting you know it is still alive. if somebody does write a
little program to parse the output, you could post it here as a
comment, that would be great.
i run the script every week, so in the worst case it could
operate for 7 days with a dead drive, and the chance is very slim that
the other drive will fail within this time.
© Copyright 2017 Tim Mackey