fixes #210: Issues with advanced_configuration section on mongodbatlas_cluster

[gmlp]

fixes #210 Issues with advanced_configuration section on mongodbatlas_cluster Opt2
This is option 2:
Based on some testing, research and some conversation with @marinsalinas and @PacoDw :

Since the migration process happens before a state change (on refresh or on apply) using the new binary. All the arguments of advanced_configuration are computed/optional meaning that these values will be retrieved from the API after state migration.

Link to any related issue(s):
closes: https://github.com/terraform-providers/terraform-provider-mongodbatlas/issues/210

Type of change:

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update
• Documentation fix/enhancement

Required Checklist:

• I have signed the MongoDB CLA
• I have read the Terraform contribution guidelines
• I have added tests that prove my fix is effective or that my feature works per HashiCorp requirements
• I have added any necessary documentation (if appropriate)
• I have run make fmt and formatted my code

Further comments

[themantissa]

Overall I think this is good - the only thing I believe we need is to add clear guidance to users about what will happen like we did for 0.4.0. Since this is the 2nd time we've needed this I'd like to do what some other providers have done and also add a new page in our Terraform doc to list information about version changes (similar to the Guides section in https://www.terraform.io/docs/providers/azurerm/index.html, with a page called Upgrade Guide for 0.6.0). That probably makes sense for another PR to create the new menu item and page/content but we'll need it before we release.

Also fyi (esp @shum before you review) - I built this tonight and ran through some options - data below. Overall it works well and as expected. As long as we explain the state migration, i.e. why advanced configs is now showing up in the plan and that a refresh w/ plan is needed I think we'll be good.
Created cluster via terraform 0.5.1, no advanced_config options in cluster resource config - state file contains:
"advanced_configuration": {
"fail_index_key_too_long": "true",
"javascript_enabled": "true",
"minimum_enabled_tls_protocol": "TLS1_2",
"no_table_scan": "false",
"oplog_size_mb": "",
"sample_refresh_interval_bi_connector": "",
"sample_size_bi_connector": ""
},

Direct API: List Adv Config API call
{
"failIndexKeyTooLong" : true,
"javascriptEnabled" : true,
"minimumEnabledTlsProtocol" : "TLS1_2",
"noTableScan" : false,
"oplogSizeMB" : null,
"sampleRefreshIntervalBIConnector" : null,
"sampleSizeBIConnector" : null

Run plan with new provider from PR #238 - Same as earlier state

Make change in .tf M10 to M20, run plan ...

Run plan without refresh:
No advanced_config change noted, M20 change found
~ provider_instance_size_name = "M10" -> "M20"

Run plan with refresh:
M20 changed found and advanced_config block now shows on output with zeros and not empty strings:
~ provider_instance_size_name = "M10" -> "M20"
advanced_configuration {
fail_index_key_too_long = true
javascript_enabled = true
minimum_enabled_tls_protocol = "TLS1_2"
no_table_scan = false
oplog_size_mb = 0
sample_refresh_interval_bi_connector = 0
sample_size_bi_connector = 0

Apply w/ refresh and now new state file is correct and no changes on the Atlas side when direct API call made:
"advanced_configuration": [
{
"fail_index_key_too_long": true,
"javascript_enabled": true,
"minimum_enabled_tls_protocol": "TLS1_2",
"no_table_scan": false,
"oplog_size_mb": 0,
"sample_refresh_interval_bi_connector": 0,
"sample_size_bi_connector": 0
}

Oplog/BI Connector changes:

Change oplog in UI ->
Terrraform plan did not pick up any change, state did not change
Terraform refresh did pick up change, changes in state file match UI change

Then set bi connector enabled in configuration and set 3 options:

advanced_configuration {
oplog_size_mb = 5500
sample_refresh_interval_bi_connector = 300
sample_size_bi_connector = 30000
}
Plan shows
advanced_configuration {
fail_index_key_too_long = true
javascript_enabled = true
minimum_enabled_tls_protocol = "TLS1_2"
no_table_scan = false
~ oplog_size_mb = 10000 -> 5500
~ sample_refresh_interval_bi_connector = 0 -> 300
~ sample_size_bi_connector = 0 -> 30000
}
Apply worked and UI matched.

[themantissa]

See comment but as long as @shum and one DoU dev approves I'm good.

[PacoDw]

LGTM!

[marinsalinas]

LGTM! great work :)