VM down - Machine Status "REPAIRING"

[Bruno M]

HI,

Yesterday my VM went down and when I accessed the console I came across this message:

NAME ZONE MACHINE_TYPE PREEMPTIBLE INTERNAL_IP EXTERNAL_IP STATUS
XXXXX-01 southamerica-east1-a custom (4 vCPU, 5.00 GiB) XXXXXX XXXXXXXX REPAIRING

I consult the google docs:

REPAIRING - The instance is being repaired. This can happen because the instance encountered an internal error. During this time, the instance is unusable. If repair is successful, the instance returns to one of the above states.

This process took about 1 hour and depois the VM dont start. The console show message "booting from hard disk 0"

Insisting with reboot, the VM started correctly. I wonder what exactly happened and prevent it from happening again.

Thanks.

[Anthony (Google Cloud Support)]

Hey Bruno, 

Thanks for your post.

In regards to your inquiry, issues such as this are internal errors at the operating system level which are out of your control. Although occurrences like this are rare, I can say this is a transient issue.

I hope this helps

[Bruno M]

Hi,

Thanks for the answer.

I searched the operating system logs (dmesg, messages centos 7), but I found nothing.

I really didn't understand what happened and I don't know how to prevent it.

[Alexandre Duval-Cid]

Hello Bruno,

Issues such as this are on the GCP side, are very rare, and resolve themselves. There's nothing you can do to prevent them except keeping a resilient architecture.

here's an article with some good practices on building a scalable and resilient architecture [1].

I hope this helps.

[1] https://cloud.google.com/solutions/scalable-and-resilient-apps