aws_workspace_workspaces plan causes intermittent "RequestError: send request failed, read: connection reset by peer"

[dgershman]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request
• If you are interested in working on this issue or have submitted a pull request, please leave a comment

Terraform Version

Terraform v0.12.26
+ provider.aws v2.67.0

Affected Resource(s)

• aws_workspaces_workspace

Terraform Configuration Files

locals {
  now = formatdate("YYYY-MM-DD HH:MM:ss", timestamp())
}

resource aws_workspaces_directory workspaces_directory {
  directory_id = "<redacted>"
  subnet_ids = [
    aws_subnet.workspaces_private_subnet_a.id,
    aws_subnet.workspaces_private_subnet_b.id
  ]

  self_service_permissions {
    change_compute_type  = false
    increase_volume_size = false
    rebuild_workspace    = false
    restart_workspace    = true
    switch_running_mode  = false
  }
}

data aws_workspaces_bundle workspaces_bundle {
  bundle_id = "<redacted>"
}

resource aws_workspaces_workspace workspace_aws_linux {
  for_each     = var.usernames
  directory_id = aws_workspaces_directory.workspaces_directory.id
  bundle_id    = data.aws_workspaces_bundle.workspaces_bundle.id
  user_name    = each.value

  root_volume_encryption_enabled = true
  user_volume_encryption_enabled = true
  volume_encryption_key          = "<redacted>"

  workspace_properties {
    running_mode                              = "AUTO_STOP"
    running_mode_auto_stop_timeout_in_minutes = 60
    compute_type_name                         = "POWER"
    root_volume_size_gib                      = 175
    user_volume_size_gib                      = 100
  }

  tags = {
    User = each.value
    Name = each.value
    FirstSeen = local.now
  }

  lifecycle {
    ignore_changes = [
      tags,
    ]
  }
}

variable usernames {
  type = set(string)
}

I have a terraform.tfvars.json file with 350 or so "usernames".

Debug Output

Panic Output

N/A

Expected Behavior

The terraform plan should have produced a plan.

Actual Behavior

Error: error listing tags: RequestError: send request failed
caused by: Post https://workspaces..amazonaws.com/: read tcp y.y.y.y:00000->x.x.x.x:443: read: connection reset by peer
Error: error reading workspace ("redacted"): RequestError: send request failed
caused by: Post https://workspaces..amazonaws.com/: read tcp y.y.y.y:00000->x.x.x.x:443: read: connection reset by peer
Error: error reading workspace ("redacted"): RequestError: send request failed
caused by: Post https://workspaces..amazonaws.com/: read tcp y.y.y.y:00000->x.x.x.x:443: read: connection reset by peer
ERROR: Job failed: exit code 1

Steps to Reproduce

1. terraform plan

The error doesn't always occur, and I believe it might be related to some kind of rate limiting panic.

Additional notes:

• Tried using a VPC endpoint, did not improve anything.
• Have tried reducing -parallelism=1
• Have tried increasing retries to 50 in the provider configuration.
• If I run the plan with -refresh=false the issue does not occur.

Important Factoids

• Running in GovCloud (US)

References

• #0000

[checker21]

I got a similar issue, when deploying a cloudfront distribution. In my setup I used S3 backend with encrypted S3 bucket. Once, I turned of encryption for S3 bucket, the error disappeared.

[Tensho]

@radius314 Hi, seems like the issue with the network, not the provider itself. However, I sincerely ask you to verify the provided config with the latest Terrfarorm and AWS provider versions 🙇

Also, I noticed you set create POWER workspaces, which usually takes a lot of time to spin up from AWS service side. If you get intermittent timeout errors, consider tweaking the recently added workspace timeouts to give the provider a little bit more time.

[lagrianitis]

Isn't this similar behaviour as #10715 ?

[github-actions]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.