[ 
https://issues.apache.org/jira/browse/CLOUDSTACK-7073?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14258444#comment-14258444
 ] 

Prachi Damle commented on CLOUDSTACK-7073:
------------------------------------------

>>>1) "So the databases prior to 4.5, which allowed the duplicate usernames 
>>>will fail to upgrade to 4.5 due to this schema change."

>>>We NEVER allowed duplicated usernames in the code, you can see the 
>>>restriction on that from the very beginning; if it happened, it was result 
>>>of this very same bug "Account/User creation: able to create user with the 
>>>same name in the same domain in Clustered MS setup". The correct fix for the 
>>>bug you've fixed, would be removing the duplicated records, and then making 
>>>the upgrade.


Yes we have allowed DB to have duplicated usernames, since there is never a 
unique constraint added on that field. From functionality perspective, we can 
delete a user and create another with same username - so in the DB, two entries 
with same username exist.

So far, the code has been checking that there is not a user with same name 
already preventing this usecase in normal scenarios. Except it breaks for the 
clustered management setup, so we need some fix.


>>>2) Your other comment "Also, traditionally we have the 'removed' column in 
>>>the DB to mark a user as removed, and no unique constraint allowing us to 
>>>add another user with same name.". >>We should file a different bug for that 
>>>- on removed event we should simply NULL the userName field in cloud.users 
>>>table. It should be done by overriding remove() call in UserDaoImpl class.

>>>There is no other way to fix this bug but by introducing the domain_id field 
>>>for unique constraint check. We can't do it in the code as it won't work in 
>>>cluster management setup. We should put back the original fix, provide the 
>>>customer with the steps to fix their environments, and fix #2

We cannot just put back the original fix without having fixes in other areas:
a) Upgrade step is needed to take care of existing customer schemas. Without 
that step this fix is not complete.
b) In the upgrade and in the API code  we need to decide on what to set the 
username field to for the removed users, when we have to introduce these 
constraints. Simply setting it to NULL is erasing the user history. Also this 
needs to be evaluated how it affects the usage functionality. Will setting it 
simply to null affect any usage record collection?
c) There is another way to fix this using the DB constraints - by having 
another column say 'inactive' tied up with the username and domainid. What this 
solves is the  limitation of the 'removed' column not being useful since it can 
have a null value and MySQL unique constraint wont work with null. Ofcourse, 
upgrade step is needed here for existing schemas.
d) Also this can be fixed through the code by having a global lock when we 
check if the username is already existing during user creation. This will solve 
the issue with cluster management setup, No?

So without these fixes we could not carry the original fix hence the revert for 
the 4.5 release.

> Account/User creation: able to create user with the same name in the same 
> domain in Clustered MS setup
> ------------------------------------------------------------------------------------------------------
>
>                 Key: CLOUDSTACK-7073
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-7073
>             Project: CloudStack
>          Issue Type: Bug
>      Security Level: Public(Anyone can view this level - this is the 
> default.) 
>          Components: Management Server
>    Affects Versions: 4.5.0
>            Reporter: Alena Prokharchyk
>            Assignee: Prachi Damle
>             Fix For: 4.6.0
>
>
> In the Java code we prohibit user to have duplicated names inside the same 
> domain. But in the DB the constraint is missing in cloud.account/cloud.user 
> table, so it is still possible to violate the rule by initiating the create 
> call from parallel threads issued either by the same MS, or by multiple MS in 
> the clustered MS setup.
> To fix, have to introduce some kind of the global lock, or db constraint 
> preventing multiple threads to insert the record with the same username.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to