3

Instance group attempting to create an instance, but can't, with this error:

The zone 'projects/myproject/zones/us-central1-f' does not have enough resources available to fulfill the request

How do I fix this, and is it something I should be expecting from GCE on regular basis?

Thanks

MeLight
  • 5,454
  • 4
  • 43
  • 67
  • I also have not been able to deploy to us-central for the last 24+ hours... – Geige V Jan 29 '19 at 02:31
  • this post might help https://stackoverflow.com/questions/52684656/the-zone-does-not-have-enough-resources-available-to-fulfill-the-request-the-re/57433015#57433015 – Yev Guyduy May 30 '20 at 14:28

1 Answers1

-3

In some rare occasions, some zones might not have enough resources available to fullfill a request. This is done to ensure that there are enough resources in each zone, to ensure that already installed users have enough of them to keep running their applications.

This type of issues are immediately noticed, and currently this is being investigated. For the moment, you can try to do one the following points:

  1. Keep trying to do the deployment until the zone has enough resources.

  2. Relax the requirements of the instance you are creating ( i.e. less CPU/Disk/Memory...)

  3. Try to deploy to another zone within the same region, for example deploy to us-central1-a. You can see the full list of available zones/regions in this documentation

I would recommend you to go for the third option, as you will be able to create the instances immediately, with the resources you need, and you probably won't be affected by the zone change.

Joan Grau Noël
  • 3,084
  • 12
  • 21
  • Hi Joan, I'm deploying a Kubernetes cluster. Can clusters in both zones be balanced by the same balancer? – MeLight Jan 29 '19 at 11:00
  • As per [this documentation](https://cloud.google.com/kubernetes-engine/docs/concepts/cluster-autoscaler#balancing_across_zones), the Load Balancer does distribute traffic between zones before autoscaling, you can check [this documentation](https://cloud.google.com/load-balancing/docs/load-balancing-overview) for details. – Joan Grau Noël Jan 29 '19 at 15:23
  • 1
    I would also consider looking at Regional Clusters, which will give you resilience against single-zone issues: https://cloud.google.com/kubernetes-engine/docs/concepts/regional-clusters – elithrar Jan 30 '19 at 02:08
  • 7
    this doesn't seem "rare" at all, I've gotten it multiple nights in a row for multiple zones in us-central1. and this is for n1-standard-1 instances so I can't really relax requirements any more. – cagenut Feb 06 '19 at 02:44