91

I am trying to delete AWS VPC (a non default one). I am getting error that "We could not delete the following VPC (vpc-xxxxxxx (xx.xx.xx.x/16)) Network interface 'eni-xxxxxx' is currently in use. (Service: AmazonEC2; Status Code: 400; Error Code: InvalidParameterValue; Request ID: xxxxxx-dd86-47c8-98cd-xxxxxxxxxxx)".

When I tried deleting the related Network Interface, I got error "eni-xxxxxxx: You do not have permission to access the specified resource."

There are security groups associated with the NI (network interface). There are subnets associated with the NI. There are no EC2 instances - all are terminated. There are no VPC endpoints. There are no RDS db, db security group or subnets.

For EC2 SG (security groups), I am getting permission denied error message while trying to delete (even the non default SG against the non-default VPC. This SG has description of "AWS created security group for d-xxxxxxxxxx directory controllers".

For subnets, I am not able to delete them because they are associated with the above mentioned NI.

Can someone help? I have a free tier account and hence even dont have technical support or neither able to post new thread in aws forums (getting error message - Your account is not ready for posting messages yet. Please try again later). I fear this is because I have a free tier account.

Ivan Aracki
  • 4,861
  • 11
  • 59
  • 73
Sujoy
  • 1,186
  • 1
  • 9
  • 12
  • Are you using Amazon Workspaces? It creates a VPC with an ENI, which you cannot delete if Workspaces has been provisioned. Are there any other services linked into your VPC, such as peering? – John Rotenstein May 15 '16 at 01:13
  • Thanks John! I am neither using Amazon Workspaces nor did establish any peering for that VPC. – Sujoy May 15 '16 at 07:48

26 Answers26

103

Check 2 things.

  • If you have a VPC Elastic IP Address you created in the VPC you are trying to delete.
  • Check if you created a AWS ELB in the VPC.

Delete above two resources if present then you will be able to delete the VPC.

Piyush Patil
  • 14,512
  • 6
  • 35
  • 54
  • thanks for reminding those two. I missed them while posting the question. I had created EIP (elastic IP) but deleted them while deleting the EC2 instance some time back. – Sujoy May 15 '16 at 07:41
  • And I had not created ELB against the VPC. So is there anything else I should check? – Sujoy May 15 '16 at 07:42
  • 3
    @Sujoy Try to detach the Internet Getaway and then try to delete the VPC. – Piyush Patil May 15 '16 at 12:54
  • Not sure of the issue. But the VPC was deleted next time I logged in after 2-3 days. Thank you all for the tips! – Sujoy Nov 05 '16 at 06:03
  • 3
    For me it was a lingering RDS instance. – Karim Tabet Dec 19 '16 at 14:42
  • Thank you. This saved me from a CloudFormation stack generated by a program that failed to delete it and then obscure cascading delete failures all the way down to the ELB without making it obvious that the ELB was the cause. – Taylor D. Edmiston Aug 22 '19 at 20:53
  • Can also be a load balancer that was not deleted. – Harel Farkash Mar 21 '22 at 06:25
  • Another place to look is the Description field of the ENI itself. For me it pinpointed to a Lambda function configured with a VPC. This data seems some important, it should be displayed on the "ENI currently in-use" pop-up screen. – DBtake3 Aug 25 '22 at 19:24
  • For me it was a NAT Gateway – Daniel Widdis Jul 31 '23 at 16:22
18

As @karimtabet said, the VPC I was unable to delete was still assigned to an active RDS instance. after removing this RDS instance, I was able to delete the VPC.

Community
  • 1
  • 1
Pierre-Yves Savard
  • 225
  • 2
  • 5
  • Something I encountered which others may run into is that Neptune also counts as an RDS when it comes to this. – SidF Oct 15 '19 at 14:22
  • This saved me!! Ty. I couldn't figure out why i couldnt delete my vpc, and i checked every other r esource i had. The only error I got during deletion was that the vpc was being used by a network interface that i couldnt delete. – dko Nov 22 '21 at 07:58
17

Check if you have any EFS instances or S3 buckets with versioning turned on or cross-domain replication (this is the one that got me recently).

The Camster
  • 1,850
  • 1
  • 20
  • 27
16

To be able to delete Security Groups (SG) and their Network Interfaces (NI) I had to:

  • delete the inbound + outbound rules of the SGs
  • wait 5-10 minutes that NIs move from state "in use" to "available"
  • delete the NIs
  • delete the SGs
Stéphane Bruckert
  • 21,706
  • 14
  • 92
  • 130
  • 1
    Worked for me as well. Had to delete all the rules and then NI and Subnet got deleted. – Aditya Singh Parmar Jun 17 '22 at 13:01
  • 2
    Here's why this worked for me: The network interface had 2 sg's associated with it. Those 2 sg's for some reason referenced each other in the inbound/outbound rules (for some unknown reason). Therefore, each individual one could not be deleted bc it contained a reference. Removed those rules, deleted sg's, deleted NIs. – Joshua Wolff Dec 09 '22 at 02:09
7

I solved this by deleting the active load balancer that was attached to the ENIs. You can check active load balancers on EC2 > Load Balancing > Load Balancers

Pedro Martins
  • 110
  • 2
  • 8
5

I was only able to delete the eni through de api:

aws ec2 delete-network-interface --network-interface-id eni-XXXXXXX

In my case I have created an ECS cluster with Terraform. But wasn't able to clean it up with the command terraform destroy.

In the web console I always got the error Network interface is currently in use.

Alexandre Ito
  • 119
  • 2
  • 2
2

A running ECS task could be an issue, try to stop/delete the same. https://forums.aws.amazon.com/thread.jspa?threadID=114301

Aruni Mishra
  • 79
  • 8
2

In My case first we have check that any load balancer available if it is then first we have to delete the load balancer then we the delete the exisiting vpc

SwissCodeMen
  • 4,222
  • 8
  • 24
  • 34
sharukh shaikh
  • 21
  • 1
1

That network interface could be connected to a security group which is then connected to another Amazon product like ElastiCache (or anything). It's connected to something which may be connected to something else.

Sam Malayek
  • 3,595
  • 3
  • 30
  • 46
1

In my case I had to first detach and delete 2 things in EC2 section (of the same region as your VPC):

  1. Network & Security -> Elastic IPs. Delete the Elastic IP associated with your VPC
  2. Network & Security -> Network Interfaces. Detach your VPC's network interface first (This took around a minute). Then delete.

Now try deleting the VPC.

rahuljain1311
  • 1,822
  • 19
  • 20
1

So you should check the network interfaces usage, if they don't appear it means you have to check computing and load balancing stacks for it.

For me it was a lambda test function that was using the interface and not appearing in the networkInterface for some reason. Make sure to check your lambda functions.

mohRamadan
  • 571
  • 6
  • 15
1

You have to do it bottom-up approach.

  1. Delete services that are working on subnets if you have them (Redshift clusters, RDS)
  2. Check jobs in Glue, Lambda etc and delete them if you don't need them
  3. Release IP addresses from NAT Gateway
  4. Detach/Delete NAT and Internet Gateway
  5. Detach and delete subnets
  6. Delete VPC
Hrvoje
  • 13,566
  • 7
  • 90
  • 104
1

Make sure to also check your lambda versions. I had deployed my application with the serverless framework and changed the VPC for deployment at some point in time. Hence I had to delete all previous versions of my lambda function to be able to delete the old VPC.

You can do that in the AWS console under the "version" tab for your lambda function.

beckersense
  • 310
  • 2
  • 12
1

If it was created by CloudFormation, deleting the CloudFormation Stack might take help you delete the VPC.

jeffhale
  • 3,759
  • 7
  • 40
  • 56
1

If you have any Endpoints configured (Currently located in the VPC section in the AWS console), you must delete these first.

Aaron McDevitt
  • 11
  • 1
1

Deleting a NAT gateway solved the issue for me.

John Mwangi
  • 11
  • 1
  • Your answer could be improved with additional supporting information. Please [edit] to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers [in the help center](/help/how-to-answer). – Community Jan 27 '23 at 08:10
0

You were right..It was ECS issue in my case. Not letting me delete my ENIs saying I dont have the permissions. I had already deregistered my ECS task definition but still It was not letting me delete resources like VPC, ENIs created by ECS.

I had to update the version of task definition under ECS using same resources as previous definition version and then I deleted that new service and then deleted the clustere.

With clustere, it deleted all resource like VPC, ENIs which it was not letting me delete earlier.

I then dergistered second version of task definition. I had two inactive tasks definitions for a monent and then one task definitions disappeared..now only one is there in inactive state.

Atul Pattewar
  • 1
0

I had forgotten that I had an RDS instance that was still connected to that VPC. After deleting it, the network interface deleted itself.

ruohola
  • 21,987
  • 6
  • 62
  • 97
0

There is a tool that can help you figure out which AWS resource associate with the ENI. This post document how you use it: https://aws.amazon.com/premiumsupport/knowledge-center/lambda-eni-find-delete/

chuan qin
  • 31
  • 1
0

I fixed this issue by deleting ElastiCache resource. After that, I was able to remove VPC

Felix Labayen
  • 385
  • 3
  • 8
0

This happened to me as well recently, and @Pierre-Yves Savard's answer helped me to understand it. Let me explain why it happens.

When you have created the whole stack with Cloudformation script, let's call it Server-Network-Stack1. Although later you created some resources ( for example RDS database ) manually from AWS Console and put them inside the Subnet/VPC which your Server-Network-Stack1 cloudformation script created previously.

Now your Server-Network-Stack1 Cloudformation stack doesn't know anything about your RDS creation, because you created it manually. So, when you are trying to remove the Server-Network-Stack1 , it will give an error means that "there is another resource which is dependent on the VPC/Subnet ( indirectly Network Interface ), and it is not declared inside my script"

Therefore, I recommended you write all of the resource declarations inside Cloudformation, even if you will create them once and will not update them throughout the project lifecycle. Otherwise, your whole team must remember which resources are created with the script, and which are not.

Ramin Mammadzada
  • 317
  • 3
  • 8
0

In my case, i have deleted the global celerator. Search for Global acelerator service and delete it.(You need AWS CLI installed and configured) Sorry my english

Carlos Vilches
  • 1
  • 1
  • Your answer could be improved with additional supporting information. Please [edit] to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers [in the help center](/help/how-to-answer). – Community Oct 19 '22 at 08:38
0

I was having trouble deleting a VPC that was associated with an RDS instance, even after the instance was deleted. I navigated to RDS->Subnet groups in the console and realised there were groups associated with the deleted RDS. By deleting them, I was then able to complete the RDS VPC deletion.

rnoodle
  • 534
  • 2
  • 5
  • 21
0

Deleting a Network Firewall did the trick for my colleague and I. Make sure your Network Firewall has deletion protection and logging disabled too

blim747
  • 51
  • 4
-1

I had no instances attached. The network interface just was related to a subnet, security group and the VPC. This is how I solved it:

  1. Delete the Load Balancers that I had
  2. Delete the network interface
  3. Delete the VPC
Alvaro Andres Pinzon Cortes
  • 1
-1

You can use this script to find your resources (nano describe_vpc.sh): When you find them, delete them one by one manually: (You need AWS CLI installed and configured) Sorry for my english :(

#!/bin/bash vpc="vpc-XXXXXXXXXXXXX" aws ec2 describe-internet-gateways --filters 'Name=attachment.vpc-id,Values='$vpc | grep InternetGatewayId aws ec2 describe-subnets --filters 'Name=vpc-id,Values='$vpc | grep SubnetId aws ec2 describe-route-tables --filters 'Name=vpc-id,Values='$vpc | grep RouteTableId aws ec2 describe-network-acls --filters 'Name=vpc-id,Values='$vpc | grep NetworkAclId aws ec2 describe-vpc-peering-connections --filters 'Name=requester-vpc-info.vpc-id,Values='$vpc | grep VpcPeeringConnectionId aws ec2 describe-vpc-endpoints --filters 'Name=vpc-id,Values='$vpc | grep VpcEndpointId aws ec2 describe-nat-gateways --filter 'Name=vpc-id,Values='$vpc | grep NatGatewayId aws ec2 describe-security-groups --filters 'Name=vpc-id,Values='$vpc | grep GroupId aws ec2 describe-instances --filters 'Name=vpc-id,Values='$vpc | grep InstanceId aws ec2 describe-vpn-gateways --filters 'Name=attachment.vpc-id,Values='$vpc | grep VpnGatewayId aws ec2 describe-network-interfaces --filters 'Name=vpc-id,Values='$vpc | grep NetworkInterfaceId

region="yy-yyyy-y" aws ec2 describe-vpc-peering-connections --region $region --filters 'Name=requester-vpc-info.vpc-id,Values='$vpc | grep VpcPeeringCo> aws ec2 describe-nat-gateways --region $region --filter 'Name=vpc-id,Values='$vpc | grep NatGatewayId aws ec2 describe-instances --region $region --filters 'Name=vpc-id,Values='$vpc | grep InstanceId aws ec2 describe-vpn-gateways --region $region --filters 'Name=attachment.vpc-id,Values='$vpc | grep VpnGatewayId aws ec2 describe-local-gateway-route-table-vpc-associations --region $region --filters Name=vpc-id,Values=$vpc | grep LocalGatewayRo>

Carlos Vilches
  • 1
  • 1
  • Your answer could be improved by formatting the commands using code blocks, and then explain what these commands do since it's a very long chain of commands. – Jan Oct 19 '22 at 08:41