New check for SQL logins with blank or Login as password

[PsPsam]

New Check

This is a new check/question
Idea taken from SQL Assessment in azure
I would like a test for logins with no password and password same as username

What would you like to check?

The test should return the count of the result. So if all is good 0 would be returned.

---Detect SQL Logins with weak passwords

SELECT count(name)
FROM master.sys.sql_logins
WHERE PWDCOMPARE(name,password_hash)=1
GO

and

---Detect SQL Logins with blank passwords

SELECT count(name)
FROM master.sys.sql_logins
WHERE PWDCOMPARE('',password_hash)=1
GO

What should be configurable for the results of the check ?

If the test should run or not

What should be able to be excluded from being tested ?

Nothing

Additional information

I took a quick look at dbatools but saw no command that does this. So maybe it a command should be created there first for this?

[SQLDBAWithABeard]

I like this test, do you know if this T-SQL will work on all versions of SQL Server?

[PsPsam]

Its from sql 2008 and up

https://docs.microsoft.com/en-us/sql/t-sql/functions/pwdcompare-transact-sql

I changed the t-sql to better align with pester test, but you can get more information. So im not sure if this should be a dbatool test function that we than can test agains in dbachecks

The origial t-sql is

---Detect SQL Logins with weak passwords 
SELECT SERVERPROPERTY('machinename') 
AS 'Server Name', 
ISNULL(SERVERPROPERTY ('instancename'), 
SERVERPROPERTY ('machinename')) AS 'Instance Name', 
name AS 'Login With Password Equal to Login Name' 
FROM master.sys.sql_logins 
WHERE PWDCOMPARE(name,password_hash)=1 
ORDER BY name 
GO

From azure

[SQLDBAWithABeard]

I agree in principle that this can be added, we would need some logic to skip the test if the version is below 2008

[michalporeba]

I was just wondering what is the value of including this function in dbatools rather than implementing it in dbachecks? especially the proposed implementation simply executes an t-sql code? Does it introduce unnecessary coupling? Would the Test-DbaLoginPassword ever get executed from dbatools directly?

[SQLDBAWithABeard]

Yes it undoubtedly would and doing so will bring more capabilities than just checking