You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ranger.apache.org by "Ramachandran (Jira)" <ji...@apache.org> on 2022/09/02 06:40:00 UTC
[jira] [Updated] (RANGER-3900) Roles deletion Takes time in Apache Ranger when there are more users
[ https://issues.apache.org/jira/browse/RANGER-3900?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ramachandran updated RANGER-3900:
---------------------------------
Summary: Roles deletion Takes time in Apache Ranger when there are more users (was: Roles deletion Takes time in Apache Ranger)
> Roles deletion Takes time in Apache Ranger when there are more users
> --------------------------------------------------------------------
>
> Key: RANGER-3900
> URL: https://issues.apache.org/jira/browse/RANGER-3900
> Project: Ranger
> Issue Type: Improvement
> Components: Ranger
> Affects Versions: 3.0.0
> Reporter: Ramachandran
> Priority: Major
>
> *{color:#0000ff}Role Deletion Steps in Apache Ranger:{color}*{color:#0000ff} {color}
>
> {code:java}
>
> 1. Getting XXRole by roleName:
> XXRole xxRole = daoMgr.getXXRole().findByRoleName(roleName); → 1DB Read call
> 2. Ensuring role is not in policy
> Long roleRefPolicyCount = daoMgr.getXXPolicyRefRole().findRoleRefPolicyCount(roleName); → 1DB Read call
> 3. Ensuring role is not in another role
> Long roleRefRoleCount = daoMgr.getXXRoleRefRole().findRoleRefRoleCount(roleName);→ 1DB Read call
> 4. Fetching RangerRole by id
> RangerRole role = roleService.read(xxRole.getId()); → 1DB Read call
> 5. Dereferencing all the users from the role.
> for (XXRoleRefUser xxRoleRefUser : xRoleUserDao.findByRoleId(roleId)) { → 1DB Read call
> xRoleUserDao.remove(xxRoleRefUser); → 1DB Write call for each xxRoleRefUser
> }
> 6. Dereferencing all the groups from the role
> for (XXRoleRefGroup xxRoleRefGroup : xRoleGroupDao.findByRoleId(roleId)) { → 1DB Read call
> xRoleGroupDao.remove(xxRoleRefGroup); → 1DB Write call for each xxRoleRefGroup
> }
> 7. Dereferencing all other roles from the role.
> for (XXRoleRefRole xxRoleRefRole : xRoleRoleDao.findByRoleId(roleId)) {→ 1DB Read call
> xRoleRoleDao.remove(xxRoleRefRole);→ 1DB Write call for each xxRoleRefRole
> }
> 8. Delete the rangerRole in DB
> roleService.delete(role); → 1DB Write call
> 9. Create TrxLog
> bizUtil.createTrxLog(trxLogList) → 1DB Write call
>
> {code}
>
> *{color:#0000ff}Total number of DB calls involved for the below role deletion in Apache Ranger:{color}*
> *{color:#0000ff}{color:#1d1c1d}Deleting the role which contain more users:{color}{color}*
> 1.role contains 100 users
> ||DB Write count||DB Read count||DB Batch Write count||Time taken to delete the role||
> |102|6|0|1.01 seconds|
>
> 1 role contains 500 users
> ||DB Write count||DB Read count||DB Batch Write count||Time taken to delete the role||
> |502|6|0| 2.241 seconds|
> 1 role contains 1000 users
> ||DB Write count||DB Read count||DB Batch Write count||Time taken to delete the role||
> |1002|6|0|3.14 seconds|
>
> {color:#0000ff}{color:#1d1c1d}*Deleting the user (Which was associated with the role .Now that role got deleted)*{color}{color}
> ||Number of users ||Time taken to delete the each user||
> |1000| 0.8 to 1.5 seconds|
> {color:#0000ff}{color:#1d1c1d}*Deleting the user (Which was not associated with any role)*{color}{color}
> ||Number of users ||Time taken to delete the each user||
> |1000| 0.8 to 1.2 seconds|
> {color:#ff0000}Recommendation :{color}
> When we reduce the number of users added into the role -→ Total number of DB calls will be reduced
> Instead of adding roles to the users ,can we add those users into group. And then link that group to the role ?
> if we use groups instead of directly adding users into role, total number DB write call will be reduced to 2 from 1002 for 1k users
>
> {color:#FF0000}Proposal{color}:
> {color:#ff0000}We can try DB batch write call instead of 1 DB write call for every user deference
> {color}
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)