Cache is generating in wrong way

[headincl0ud]

I'm using Terragrunt to manage a quite large amount of AWS accounts (>100).

I don't need to categorize these accounts per region/env (every account is independent, managed by different unit), my current directory strategy is looking in this way:

├──cloud-configuration
│   ├── infra
│   ├── terragrunt.hcl
│   │   ├── IT
│   │   │   ├──1234567890
│   │   │   │   ├── ap-northeast-1
│   │   │   │   ├── ap-northeast-2
...
│   │   │   │   ├── ap-south-1
│   │   │   │   │   ├── region.hcl
│   │   │   │   │   └── CONFIG_RULES
│   │   │   │   │       └── REQUIRED_TAGS
│   │   │   │    │   │       └── terragrunt.hcl
│   │   │   │   ├── us-west-1
│   │   │   │   └── us-west-2
│   │   │   │   ├──2345678901
│   │   │   │   ├── ap-northeast-1
│   │   │   │   ├── ap-northeast-2
...
│   │   │   │   ├── ap-south-1
│   │   │   │   │   ├── region.hcl
│   │   │   │   │   └── CONFIG_RULES
│   │   │   │   │       └── REQUIRED_TAGS
│   │   │   │    │   │       └── terragrunt.hcl
│   │   │   │   ├── us-west-1
│   │   │   │   └── us-west-2
├── cloud-modules-aws-config-rules
│   ├── REQUIRED_TAGS

As the administrator, I'm implementing unique account configurations, like AWS Config Rules. Every region has a different configuration (different pack of AWS Config Rules).

Unfortunately, I realized that Terragrunt is generating for every region and for every rule an incredible size of the cache.

Issue:

├──cloud-configuration
│   ├── infra
│   ├── terragrunt.hcl
│   │   ├── IT
│   │   │   ├──1234567890
│   │   │   │   ├── ap-northeast-1
│   │   │   │   ├── ap-northeast-2
...
│   │   │   │   ├── ap-south-1
│   │   │   │   │   ├── region.hcl
│   │   │   │   │   └── CONFIG_RULES
│   │   │   │   │       └── REQUIRED_TAGS
│   │   │   │    │   │       └── terragrunt.hcl
│   │   │   │    │   │       └── .terragrunt-cache
│   │   │   │    │   │   │          └──cloud-configuration
│   │   │   │    │   │   │   │          └──infra 
│   │   │   │    │   │   │   │          └──... 
│   │   │   │   ├── us-west-1
│   │   │   │   └── us-west-2
│   │   │   │   ├──2345678901
│   │   │   │   ├── ap-northeast-1
│   │   │   │   ├── ap-northeast-2
...
│   │   │   │   ├── ap-south-1
│   │   │   │   │   ├── region.hcl
│   │   │   │   │   └── CONFIG_RULES
│   │   │   │   │       └── REQUIRED_TAGS
│   │   │   │    │   │       └── terragrunt.hcl
│   │   │   │    │   │       └── .terragrunt-cache
│   │   │   │    │   │   │          └──cloud-configuration
│   │   │   │    │   │   │   │          └──infra 
│   │   │   │    │   │   │   │          └──... 
│   │   │   │   ├── us-west-1
│   │   │   │   └── us-west-2
├── cloud-modules-aws-config-rules
│   ├── REQUIRED_TAGS

For every account (>100) the cache is generated to every region (16) and for every module (every AWS Config Rule) In the end, the execution took hours to plan all and apply all.

[headincl0ud]

https://github.com/gruntwork-io/terragrunt/issues/1390

[denis256]

Hi, I was wondering if can be used --terragrunt-download-dir <persistent-path> to help keep cached dependencies between runs

[headincl0ud]

Hi @denis256 ,

The --terragrunt-download-dir is not the solution in this case. The Terragrunt will create the same amount of cache and multiplied directories but in different locations.

[igkins]

@headincl0ud you come up with a solution? we're facing something similar

[gairik-sablono]

Same issue here. As we add more and more resources, terragrunt seems have to become slower and slower.

[ashuraits]

+1 very slow

[kahootali]

Extremely slow for me as well, anyone found a solution?