LDAP Constraint Violation When Changing Password in AD through ldapmodify

Question

I currently try to change passwords in our Active Directory Envoirenment via LDAP on Linux since the users in question do not have access to a windows-machine and we want to keep it that way. In order to change the password I am currently stuck figuring out how to use ldapmodify to do so. After a lot of reading on different sites/forums/newsgroups I am much more confused than before

However: I try the following command to do so:

ldapmodify -f ldif.example -H ldaps://lab01-dc01.example.com -D 'CN=test,CN=users,DC=lab01,DC=example,DC=com' -x -W

The contents of the ldif.example:

dn: CN=test,CN=Users,DC=lab01,DC=example,DC=com
changetype: modify
delete: unicodePwd
unicodePwd:: V3VQdXV1STEyLg==
-
add: unicodePwd
unicodePwd:: QmxhVVVraTEyLg==
-

(Don't worry - those passwords are not used anywhere and it is not a production envoirenment)

Now - every time I execute the command I get the following error:

modifying entry CN=test,CN=Users,DC=lab01,DC=example,DC=com"
ldapmodify: Constraint violation (19)
 additional info: 0000216C: AtrErr: DSID-03190EB0, #1:
 0: 0000216C: DSID-03190EB0, problem 1005 (CONSTRAINT_ATT_TYPE), data 0, Att 9005a (unicodePwd)

Now, after what I read the reason for this error is either that the password is badly formatted or that the password policy doesn't allow the password I used. I checked the policy - multiple times now - and the new password definetly complies to the policy by all the criteria. If I set the password using a Windows-machine it also works well (of course I changed the "oldpassword" and "newpassword" afterwards since I am not allowed by the policy to change to an earlier password). The password I enter after passing the "-W" option to ldapmodify is also definetly right, otherwise the error spit out by ldapmodify is that I used invalid credentials instead of a constraint violation. So - the sole reason I can think of is indeed a bad formatted password - but I can't figure out where the bad formatting should come from since I use the normal base64 algorythm to encode the password.

Has anyone an idea what is going on? Can anyone push me in the right direction?

Help is very appreciated and I thank you in advance.

Edit: Something which bugs me: When I run the base encoded strings through base64 it keeps telling me "Invalid Input". Now - I went ahead and just re-coded the passwords with the use of base64 on the linux machine - but when I run the generated string through the decode function again, base64 keeps telling me "Invalid Input"... The strings however slightly changed between the windows-base64 encoded string and the linux encoded string. But base64 just says "Invalid input" no matter what I put in there.

Edit2: Nevermind - reading the purpose of the function I gather that it throws this error because of the dots and the exclamation mark in the password.

Answer 1

For future reference, if anyone should encounter similiar problems: The simple solution? Just use smbpasswd instead of ldap to change the password - that works flawless! I am really grumped that I didn't think of it before :D

However - the way to change your password in the active directory using samba (using CentOS):

~#yum install samba
~#smbpasswd -r domaincontroller.example.com -U testuser1
Old SMB password:
New SMB password:
Retype new SMB password:
Password changed for user testuser1 on domaincontroller.example.com

And then you can login using the new password. Easy as that, really.