I have a bash script which activate multiples configurations with following gcloud command:

gcloud config configurations activate MYENV

In the most case, this command works instantly, sometimes it takes 10 secondes or more than 60 secondes.

I am using sleep(10) ... but it is not the best way. Anyone, have an idea on how to wait that gcloud command finish?

Even if I can use return=$? , I will not know if this command waits for the command to be finished.

CodePudding user response：

You can use the wait command to use for the last command identified by its PID:

#!/bin/usr/env bash

# launch your command in the background
gcloud config configurations activate MYENV &

# wait for the command to finish
wait $!

echo 'Configuration $MYENV activited'

CodePudding user response：

I'm breaking a personal practice by answering a question by proposing an alternative approach but...

gcloud config configurations adjusts gcloud's global state by swapping in a set of configuration variables. These variables are all readily represented directly in gcloud commands.

Particularly when (writing) scripts, I feel a better practice is to use flags and variables on commands explicitly.

Rather than:

for CONFIG in "foo" "bar" "baz"
do
  gcloud config configurations activate ${CONFIG}
  gcloud compute instances delete instance-01 \
  --quiet
done

Where parameters (flags) of gcloud compute instances delete insurance-01 are implicit.

Versus:

for PROJECT in "foo" "bar" "baz"
do
  for ZONE in "us-west1-c" "us-central1-a"
  do
    gcloud compute instances delete instance-01 \
    --zone=${ZONE} \
    --project=${PROJECT} \
    --quiet
  done
done

Where you:

• avoid any global update "pauses"
• avoid breaking a gcloud session running in a different process
• have a more explicit and understandable set of commands
• can pump the script output to a log file to see exactly what ran

Personally, I discourage use of gcloud config for anything other than the unavoidable i.e. gcloud config ... account and that may be all.