Exercise

Collecting Disk Data Exercise

Task

Write a script that collects disk information from every host named on the command line and stores each host's results in its own file.

Hosts: servera and workstation
Script location: ~/scripts/collect_disk_data.sh
Output location: /tmp/disk-data/

Create a basic script file

Create the scripts directory, move into it, and open a new file.
```
mkdir -p ~/scripts
cd ~/scripts
vim collect_disk_data.sh
```
Start with the hashbang and a line that proves the script ran.
```
#!/bin/bash
echo "Starting $0"
```
Hand the file to bash and it runs.
```
bash collect_disk_data.sh
```
No execute permission needed yet Naming bash yourself means bash opens and reads the file like any other file. The execute bit only matters when you ask the kernel to run the file directly.
Make it executable

Add the execute bit, then run the file itself rather than handing it to bash.
```
chmod a+x collect_disk_data.sh
./collect_disk_data.sh
```
Look at $0 in the output. It changes from collect_disk_data.sh to ./collect_disk_data.sh, because $0 reports how the script was invoked.
Write a basic for loop

Replace the announcement with a loop over a fixed list of hosts. Get the loop itself working before adding anything else to the body.
```
vim collect_disk_data.sh
```
```
#!/bin/bash
echo "Starting $0"
for host in workstation servera; do
    echo $host
done
```
```
./collect_disk_data.sh
```
The body runs once per host, and the loop ends when the list runs out.
Iterate on the logic

Get one host working by hand before putting the commands in the loop, so you know what they produce.
```
mkdir -p /tmp/disk-data
ssh student@servera lsblk > /tmp/disk-data/servera-disk-data.txt
ssh student@servera df -h >> /tmp/disk-data/servera-disk-data.txt
ls -l /tmp/disk-data
cat /tmp/disk-data/servera-disk-data.txt
```
Note the redirection operators The first command uses > to create the file. The second uses >> so that it adds to the file instead of replacing what the first one wrote.

Now move those commands into the loop, using $host in place of the fixed host name so the same three lines run for every item in the list.
```
vim collect_disk_data.sh
```
```
#!/bin/bash
echo "Starting $0"
for host in workstation servera; do
    echo $host > /tmp/disk-data/$host-disk-data.txt
    ssh student@$host lsblk >> /tmp/disk-data/$host-disk-data.txt
    ssh student@$host df -h >> /tmp/disk-data/$host-disk-data.txt
done
```
```
./collect_disk_data.sh
ls -l /tmp/disk-data
cat /tmp/disk-data/workstation-disk-data.txt
```
Make the script take arguments

The host list is still hard coded, so the script can only ever inspect those two machines. Replace the list with "$@".
```
vim collect_disk_data.sh
```
```
for host in "$@"; do
```
Read the whole script back, then clear the previous results and run it again, this time naming the hosts on the command line.
```
cat collect_disk_data.sh
rm -f /tmp/disk-data/*
./collect_disk_data.sh workstation servera
ls -l /tmp/disk-data
cat /tmp/disk-data/servera-disk-data.txt
```
Quote "$@" The quotes keep each argument as one item, even if an argument contains a space. Unquoted $@ would split such an argument into two loop iterations.

Clean up

Remove the collected data and the script.

rm -rf /tmp/disk-data
rm ~/scripts/collect_disk_data.sh
cd ~

Task

Create a basic script file

Make it executable

Write a basic for loop

Iterate on the logic

Make the script take arguments

Clean up