You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
First, awesome repo @matt-FFFFFF ... Found this extremely helpful when learning how to deploy VMSS ala Bicep.
What I've come to realize is the ado pipeline will pick-up an instance/agent before cloud-init has completed the post-deployment installations. I thought about incorporating cloud-init status --wait to ensure the pipeline does not progress until cloud-init completes the post-deployment tasks, but that's pretty messy. Every pipeline would need this noise... And the next stage could pick up a totally different instance/agent (let's say one that has not completed the cloud-init).
While not directly related to this repo (everything works beautifully), have you run into this issue before? Curious what your solution was to this dilemma. Thanks for your contributions!
The text was updated successfully, but these errors were encountered:
It's been a while since I used this but I seem to remember a race condition where the cloud init installation took longer than the registration with ADO.
This was a one time issue though IIRC? Once it has provisioned then the VM sits there ready to go for next time.
Let me know if this chimes with you or if I've missed something.
First, awesome repo @matt-FFFFFF ... Found this extremely helpful when learning how to deploy VMSS ala Bicep.
What I've come to realize is the ado pipeline will pick-up an instance/agent before cloud-init has completed the post-deployment installations. I thought about incorporating cloud-init status --wait to ensure the pipeline does not progress until cloud-init completes the post-deployment tasks, but that's pretty messy. Every pipeline would need this noise... And the next stage could pick up a totally different instance/agent (let's say one that has not completed the cloud-init).
While not directly related to this repo (everything works beautifully), have you run into this issue before? Curious what your solution was to this dilemma. Thanks for your contributions!
The text was updated successfully, but these errors were encountered: