Shortage of Compute Engine Resource in us-central-1

[skycruiser8]

Hello everyone. I am currently working as a lab assistant in a university in Indonesia and are currently teaching GCP to about 450 students. We are using the Higher Education Grants system and mostly use us-central-1 region since the region looks like it has the cheapest prices available. However, as of today, we cannot acquire any resources from said region. Here is the error message:

A e2-micro VM instance is currently unavailable in the us-central1-a zone. Alternatively, you can try your request again with a different VM hardware configuration or at a later time. For more information, see the troubleshooting documentation.

Is there anything that we can do? We currently try to redirect students to other regions or machine types. However, this does add complexity to the students to setup the machines all over again. Thank you very much for any help that will come 😄

[ortegafernando]

Hi, we have also a e2-micro VM and after stoping it to reboot it.... since yesterday is impossible, We get this error message: 

Failed to start instance-1: A e2-micro VM instance is currently unavailable in the us-central1-a zone. Alternatively, you can try your request again with a different VM hardware configuration or at a later time. For more information, see the troubleshooting documentation.

Is google going to increase resources in this zone? How could I see which zones are with free resources at least to move our VM ?

Thanks.

[rmoskal]

Hmm, I was able to launch a c2-standard-8 without any problems. The e2 instances still don't work. Maybe google is pushing us towards the more expensive SKUs!
Happy Thanksgiving!

[GCPDude]

Same Problem here. Stopped an e2-micro and not able to reboot in us-central1-a zone. Tried to move to us-east1 - also not working. 

Finally, I was able to move to us-west1-b successfully. 

[ortegafernando]

Hi, how do you move? Thanks

[elceed]

Hi, unfortunately, zones are permanent when you create the instances. You can't edit it. If the VPS doesn't have much data, I say make a new instance in the West region. That worked for me.

[rmoskal]

You can create an image from your instance and then create a new instance in whatever zone you like.

[Yazz]

That's great, but moving your instance shouldn't have to be an option. This happens with GCP every few months, and I think we need to move to AWS. Good thing we're cloud native.

[rmoskal]

I've been unable  to launch a  e2-custom-8-18432 vm instance in gcloud region us-east4-c since yesterday.

I recently moved it from a to c because the vm scheduler was failing to start it successfully.

That cleared up the problem for a few weeks, but now it's started in c.

[lutian]

same problem 

[carlos-ponce]

Hello,

 I am experiencing a similar issue. I am currently working on a cloudskillsboost.google challenge.

 While trying to spin up a f1-micro instance I get the following error:

A f1-micro VM instance is currently unavailable in the us-east1-b zone. Alternatively, you can try your request again with a different VM hardware configuration or at a later time. For more information, see the troubleshooting documentation.

Changing the zone has not solved the issue.

[PromptToPrompt]

same issue

[MindsClik]

same issue we have development machines and no one is able to work, f1 micro, g1 small none of them are started .. all have same issue in us-east1-b

[VirtualMe]

Me too.....busy with a LAB where I need to raise a n1-standard-1 (1vCPU, 3.75 GB memory) but get the VM hardware unavailable......tried on 24th and 25th.....am meant to be creating in us-central1-c have tried another machine type and us-central1-a same issue.......I'm presuming it has something to do with Black Friday and maybe a temporary ban on creating vm's associated with training LABs .....but this is speculation on my part.

[MindsClik]

facing the same issue in us-east1-b ... 

[elandyg]

I had to route to Montreal yesterday... I'm able to link up to east4, but everything else from us-central to and east seems to be out of commission.

Does anyone have a good command line workflow to deploy and retry? 

[1011]

Creation of VM's on us-central1 is down as well since the 23rd. In my error logs I get  ZONE_RESOURCE_POOL_EXHAUSTED

[PC-2107]

I was able to up my instance by changing my config from e2-medium to n2-standard-2..

Hope it helps you..

[Troy]

That worked for me as well, thanks for the post!!

[DangerMouse]

This worked for me as well, but I didn't see what the pricing difference was going to be.  I have a really small program running daily M-F so don't need anything more than an E2 instance.

Several months ago I had this issue with Vertex AI for spinning up Jupyter Notebooks.  I got so frustrated, I ended up just not using Vertex AI any more.

[rmoskal]

Changing the machine type to c2-standard-8 has also resolved my problems.  I guess google is pushing us to the more expensive SKUs!

[VP_DevOps]

I'm having the same problem. Would the solution be to perform a zone migration or improve the hardware characteristics of the VM?

[1011]

Zone migration would be magical.

With the holidays this past week, most of the company off, I so happened to login on my off day and thats how I noticed one of our batch scripts stopped executing.  We have a VM which auto gets created every 6 hours and it executes a docker image. Well, it stopped working since Wednesday due to this lack of resources on us-central1. Fortunately it won't be anything super serious on our end, but what if it was? 

It would be great if there was infrastructure behind the scenes on Google's side that could do something like:

If VM creation on us-central1 fails for e2-highcpu-2 then
Attempt VM creation on us-west1 with e2-highcpu-2

[jeisson-udata]

On kubernetes services this is present, i dont to create a new cluster with pods.

 When  google could fix this error?

[CoderNick]

Simple (temporary) fix:

Step 1 : Go to instance screen, hit the 3 dots and create a new machine image. For the isntance select the instance you are looking to duplicate and for location select multiple regions. Name, Description and Encryption don't matter (at least not for me).
Step 2 : Once that is created, on instance screen, create a new image from machine instance, then give it a new region.

With that being said I'm looking for a more permanent fix here. I use an instance schedule for my instance 5 days a week (workdays). Is there way for you to add an if statement to a schedule?

For example:

Instance 1 is in US-East1

Instance 2 (backup) is in US-East4

Try to start instance 1. If instance 1 fails to start then start instance 2. This would be a good backup just in case this happens again. 

[MJGrade100]

 i am getting this error please help me 

[ccisternas]

we are experiencing the same problem.

[Syacko]

Just restarted my instance and it worked.

[MirageF1]

Same issue here, trying to start my self setup/configured 4x E2 medium (2 vCPU shared core, 4GB RAM/20GB Standard Persistent Disk)  VM Kubernetes cluster located in us-east1-b which is failing with error "A e2-medium VM instance is currently unavailable in the us-east1-b zone. Alternatively, you can try your request again with a different VM hardware configuration or at a later time. "

This is not a GKE managed cluster.

First time I've had this error in over a yr and a half of setting up my Kub cluster, but the concern is this seems to be ongoing in us-east1-b....

I don't want to go through the shlep of creating machine images and recreating the whole 4 vm cluster setup in a new zone including the load balancer and routing which was all done manually and may not even work and fail with same out of resource error! Also the pricing will be inhibitive - additionally my 4x20GB disks reside on east1-b and will also need to be copied/recreated in the migrated zone with associated network data txfer costs.

I'm running this as a test/learning environment, so $'s are the bottom line as it's paid from my pocket.

I'm not based in US so will try again in morning when hopefully US zones are quieter but I have my doubts reading others similar ongoing issue.

The cynic in me says it's unlikely to be a Black Friday special discount issue with additional resources created ?( I certainly wasn't offered any for my 18 months of loyalty :))

More likely Google wanting/nudging/trying to coerce us to move to more expensive hardware or zones.

[SilverPresident]

This is unbelievable. I never expected to have this kind of issue with google especially after so many years of no problems. I am most disappointed that there seems to be nothing coming from google about this issue. US-Central, US-East are both having this issue and if so many persons are unable to restart there machine, how come resources are still exhausted?

My system have been down for 3 days now! Not happy.

[lutian]

still no official reply from Google?

[Yazz]

They are very annoying. This happens with GCP every few months, and I'm thinking we maybe need to move to AWS.

[jesusandres]

same problem. solved it creating a new instance based on an image from my first instance. But this time in a different region. in us-west1-b.

[Yazz]

Great, but moving your instance shouldn't have to be an option. This happens with GCP every few months, and I'm starting to think GCP was the wrong choice for us.

[Ronson_Xaviour]

Looks like resources are available now. I'm able to create e2 instances.