Removal of user with dependencies fails

[jufemaiz]

Terraform Version

v0.11.3

Affected Resource(s)

Please list the resources as a list, for example:

• pagerduty_user
• pagerduty_schedule
• pagerduty_escalation_policy

Terraform Configuration Files

This one is a little hard due to the structure. In simple form, we had:

• an array of pagerduty_users
• a pagerduty_schedule that the users were members of
• two pagerduty_escalation_policies that referenced the pagerduty_schedule

Debug Output

Plan:

------------------------------------------------------------------------
  |  
  | An execution plan has been generated and is shown below.
  | Resource actions are indicated with the following symbols:
  | ~ update in-place
  | - destroy
  |  
  | Terraform will perform the following actions:
  |  
  | ~ pagerduty_escalation_policy.blah0
  | rule.1.target.#:                    "4" => "3"
  | rule.1.target.3.id:                 "XXXXXX0" => ""
  |  
  | ~ pagerduty_escalation_policy.blah1
  | rule.#:                             "2" => "1"
  | rule.1.escalation_delay_in_minutes: "30" => "0"
  | rule.1.target.#:                    "1" => "0"
  | rule.1.target.0.id:                 "XXXXXX1" => ""
  |  
  | ~ pagerduty_escalation_policy.blah2
  | rule.1.target.#:                    "4" => "3"
  | rule.1.target.3.id:                 "XXXXXX0" => ""
  |  
  | ~ pagerduty_escalation_policy.blah3
  | rule.#:                             "2" => "1"
  | rule.1.escalation_delay_in_minutes: "30" => "0"
  | rule.1.target.#:                    "1" => "0"
  | rule.1.target.0.id:                 "XXXXXX1" => ""
  |  
  | ~ module.pagerduty.pagerduty_schedule.on_call_shift
  | layer.0.users.#:                    "4" => "3"
  | layer.0.users.2:                    "XXXXXX0" => "XXXXXX2"
  | layer.0.users.3:                    "XXXXXX2" => ""
  |  
  | - module.pagerduty.pagerduty_user.devops[3]

Expected Behavior

• User is able to be destroyed when the relevant dependencies are deleted as part of the terraform plan to be applied.

Actual Behavior

• User failed to be deleted as dependencies were not destroyed first (and were part of the plan).

module.pagerduty.pagerduty_user.devops[3]: Destroying... (ID: XXXXXX0)
  |  
  | Error: Error applying plan:
  |  
  | 1 error(s) occurred:
  |  
  | * module.pagerduty.pagerduty_user.devops[3] (destroy): 1 error(s) occurred:
  |  
  | * pagerduty_user.devops.3: DELETE API call to https://api.pagerduty.com/users/XXXXXX0 failed 400 Bad Request. Code: 0, Errors: [The user cannot be deleted as they have 2 escalation policies and 1 schedule. Please remove the user from the following escalation policies and remove the user from the following schedule to continue.], Message:

[roccomuscaritolo-okta]

This is also effecting us. Having to have 20 repeated pagerduty_team_membership resources is kinda crazy. So we attempted to use count.

However, if you test removing a user from the middle of the list that terraform count iterates over, it shifts the rest of the entries down one index and forces new resources to be made for each.

And this process fails, because terraform tries to remove the user(s) from the Nth count and re-add it in Nth - 1 while it is still a member of the scheduled rotations...

pagerduty_team_membership and terraform counts don't work. We need it to be a list.

[stmcallister]

Hello! #258 should have addressed this. Please post again if that's not the case.

[jufemaiz]

Thanks for the update @stmcallister !

Sadly in the interim I've left the company that I was at when trying to make use of it and that company subsequently went under. Can't test this personally.

[evanrappe]

@stmcallister, have you heard from anyone on this issue since last year? Seem to still be experiencing it when attempting to remove a user, team membership, schedule. I haven't tested a ton of scenarios but it seems to maybe hit a snag when an escalation policy references a schedule.

...failed 400 Bad Request. Code: 2001, Errors: [User cannot be removed as they belong to an escalation policy on this team]

[frek818]

I also have this issue. Any solutions?