GitLab variable not translated in runtime - gitlab

Question

I've a GitLab Runner, connected to a real device (acting as server or similar). I'm using GitLab and my Runner to automate this device and basically remote-control it.
I want now to feed a variable to the device. In this case, an IP address, and i want to do it through GitLab variables.
The problem is that in runtime, the variable is not substituted with the actual IP address and it lands to the device in GitLab Variable format. This causes the server to send an error since it cannot understand the given value.
Is there any way i could use through GitLab variables that would allow me to give this device this parameter, being sure that the variable gets substituted with the actual parameter before being send?
Thank you
This is how i defined the parameter in the gitlab-ci file:
variables:
IPADDRESS_1: 170.x.x.10
...
The other file is the yml configuration file for my server (the actual device).
NOTICE: this is NOT the gitlab-ci file, but a server-specific yml configuration file, which is copied from GitLab to the server itself.
...
export_version: '0.2'
name: Example configuration yml
host: ${IPADDRESS_1}
...
This is the error message from the server (notice that the variable was not translated):
"errors": {
"invalid_ip": [
{
"id": null,
"name": "Example configuration yml",
"host": "${IPADDRESS_1}",
...

Answer 1

Regarding the questions itself - "GitLab variable not translated in runtime" - why would id? Gitlab runner simply executes commands in a shell of your choice, and a proper shell won't modify your files until you explicitly told it to.
The easiest way to substitute variables with values in a file using bash would be to use envsubst command. As per man:
envsubst - substitutes environment variables in shell format strings
In normal operation mode, standard input is copied to standard output, with references to environment variables of the form $VARIABLE or ${VARIABLE} being replaced with the corresponding values.
If a SHELL-FORMAT is given, only those environment variables that are referenced in SHELL-FORMAT are substituted; otherwise all environment variables references occurring in standard input are
substituted.
So in your case, assuming you've set IPADDRESS_1 to 170.x.x.10 in gitlab, and have a file called test.yaml with the following content:
"errors": {
"invalid_ip": [
{
"id": null,
"name": "Example configuration yml",
"host": "${IPADDRESS_1}",
Executing command
envsubst < test.yaml > substituted.yaml
Will create or overwrite, if you already have a file with this filename, substituted.yaml, with the following content:
"errors": {
"invalid_ip": [
{
"id": null,
"name": "Example configuration yml",
"host": "170.x.x.10",
Be aware, that you can't read and write to the same file using shell redirects (i.e. using envsubst < test.yaml > test.yaml, because doing so will just erase contents of test.yaml, see https://unix.stackexchange.com/questions/36261/can-i-read-and-write-to-the-same-file-in-linux-without-overwriting-it
To test this locally, you can use export IPADDRESS_1=170.x.x.10 before running previous command.
If you have some complex, and want to generate yaml (or any other text file), based on some variables, i suggest you to use some kind of templating engine, i.e. Jinja.