User manager in meta #400

boshengchen · 2019-05-17T07:51:59Z

Instead of PR #257 .

Addressed comments.
storage the data using dataman .

nebula-community-bot · 2019-05-17T07:52:04Z

Can one of the admins verify this patch?

src/meta/processors/AuthenticationProcessor.cpp

src/interface/meta.thrift

src/meta/MetaServiceUtils.cpp

boshengchen · 2019-05-24T02:13:59Z

Thanks for your response @darionyaphet @dangleptr , I have two questions to discuss as below :
1, About user options, In fact, my original design and implementation was based on mysql ，There are options for default roles at that time. It was changed to what it is now after a couple of iterations, Thanks for your suggestions during the design phase. About option email and phone, The implementation of the syntax has been merged into the mainline，seems that we need to get a confirmed design again . and then re-implement these confirmed design. So, Please feel free to let me know if have any good ideas . Let's using issue #190 for recording.
2, About dataman , what are some good ways to avoid using dataman in module meta? Please refer to #190 for requirements. Thanks.

boshengchen · 2019-05-27T06:47:34Z

Address some comments as below:
1, Changed user data transport protocol from dataman to thrift.
2, Removed user property email and phone.
3, Added new user property lock status, this option is referenced from mysql.

src/interface/meta.thrift

boshengchen · 2019-05-31T02:28:46Z

About SpaceID and UserID exists multiple places, I've actually thought about it，Because these two structures are used for user management often. So can use these names directly . I think convert
from name to id is superfluous in here. What do you think?
UserID and SpaceID are actually useful, they will be used for privilege management. I will use UserID and SpaceID to implement some caches in meta server or meta client.

dangleptr · 2019-05-31T02:57:14Z

About SpaceID and UserID exists multiple places, I've actually thought about it，Because these two structures are used for user management often. So can use these names directly . I think convert
from name to id is superfluous in here. What do you think?
UserID and SpaceID are actually useful, they will be used for privilege management. I will use UserID and SpaceID to implement some caches in meta server or meta client.

It make sense for userId, but not for spaceID. We'd better use spaceId to join different entries instead of spaceName.

dangleptr · 2019-05-31T05:31:43Z

By the way, please refer #483 . Put the user related processors under a separate dir.

boshengchen · 2019-05-31T07:29:03Z

It make sense for userId, but not for spaceID. We'd better use spaceId to join different entries instead of spaceName.
That makes sense, let me improve the space name and RoleItem related.

boshengchen · 2019-06-03T07:22:05Z

1,Refactor user processors code structure.
2,Remove user properties first name and last name.
3, Add new user properties of resource limit.
4, Changed related thrift structure from 'name' to 'id',
5,Add some comments for role.

boshengchen · 2019-06-03T07:27:19Z

Ready to review. Please review again. Thanks.

dangleptr · 2019-06-04T02:30:41Z

src/graph/PermissionManager.h

+#ifndef GRAPH_PERMISSIONMANAGER_H
+#define GRAPH_PERMISSIONMANAGER_H
+
+// Operation and permission define:


dangleptr · 2019-06-04T02:32:41Z

src/meta/processors/BaseProcessor.inl

@@ -162,6 +162,18 @@ Status BaseProcessor<RESP>::spaceExist(GraphSpaceID spaceId) {
 return Status::SpaceNotFound();
 }

+template<typename RESP>
+Status BaseProcessor<RESP>::userExist(UserID spaceId) {


Could we unify the userExist with getUserId ?

userExist is judged by UserID , But getUserId is judged by userName. And they query for different tables userIndex and userData. They are same with spaceExist and getSpaceId. I think it's best not to unify them. WDYT?

Fair enough.

dangleptr · 2019-06-04T02:34:35Z

src/meta/processors/partsMan/DropSpaceProcessor.cpp

@@ -41,6 +41,16 @@ void DropSpaceProcessor::process(const cpp2::DropSpaceReq& req) {
 deleteKeys.emplace_back(MetaServiceUtils::indexSpaceKey(req.get_space_name()));
 deleteKeys.emplace_back(MetaServiceUtils::spaceKey(spaceId));

+ // delete related role data.


Now we just delete the parts information when deleting space. We could delete the useless user and schema when compaction.

Yeah, that makes sense. Let me modify it later.

dangleptr · 2019-06-04T02:39:57Z

src/meta/processors/usersMan/AuthenticationProcessor.cpp

+}
+
+
+void AlterUserProcessor::process(const cpp2::AlterUserReq& req) {


It seems the alter is to change the whole value.
I want to know what's the language for alter user?

No, Only change properties that require alter.

Alter user sentences:

ALTER USER account WITH with_user_opt_list with_user_opt_item : ACCOUNT LOCK | ACCOUNT UNLOCK | MAX_QUERIES_PER_HOUR INTEGER | MAX_UPDATES_PER_HOUR INTEGER | MAX_CONNECTIONS_PER_HOUR INTEGER | MAX_USER_CONNECTIONS INTEGER ; with_user_opt_list : with_user_opt_item | with_user_opt_list COMMA with_user_opt_item ;

dangleptr · 2019-06-04T02:41:22Z

src/meta/processors/usersMan/AuthenticationProcessor.cpp

+
+
+void ChangePasswordProcessor::process(const cpp2::ChangePasswordReq& req) {
+ folly::SharedMutex::WriteHolder wHolder(LockUtils::userLock());


Change pwd is a special sentence?

Yes, Change password is a separate sentence. No need to verify the password if it is a GOD role. else the password must be validated. And GOD can change anyone's password. Other roles can only change their own password.

change_password_sentence : CHANGE PASSWORD account TO new_password | CHANGE PASSWORD account FROM old_password TO new_password ;

dangleptr · 2019-06-04T02:42:37Z

src/meta/processors/usersMan/AuthenticationProcessor.cpp

+ return;
+ }
+
+ if (!checkPassword(userRet.value(), req.get_encoded_pwd())) {


I want to know when to check the password?

This is a reserved interface，Used for client or terminal connections.

src/meta/processors/usersMan/AuthenticationProcessor.cpp

src/meta/test/CMakeLists.txt

src/meta/test/TestUtils.h

boshengchen · 2019-06-14T09:04:02Z

Address laura-ding's comment

src/meta/test/CMakeLists.txt

…moved user property email and phone. 3, Added new user property lock status

…irst name and last name. 3, Add new user properties of resource limit. 4, Changed related thrift structure from 'name' to 'id', 5,Add some comments for role.

…authentication_test module

dangleptr · 2019-06-17T02:30:16Z

Jenkins go

nebula-community-bot · 2019-06-17T03:05:53Z

Unit testing passed.

laura-ding · 2019-06-19T02:30:03Z

src/meta/test/CMakeLists.txt

@@ -171,3 +171,33 @@ nebula_link_libraries(
 gtest
 )
 nebula_add_test(meta_http_test)
+add_executable(


There are no aligned with 4 Spaces

A stupid mistake，corrected.

laura-ding

Well done

boshengchen · 2019-06-19T03:12:58Z

Jenkins go

dangleptr · 2019-06-19T06:31:05Z

Jenkins go

nebula-community-bot · 2019-06-19T06:59:46Z

Unit testing passed.

dangleptr

Well done

* User manager of Meta Server * 1, Changed user data transport protocol from dataman to thrift. 2, Removed user property email and phone. 3, Added new user property lock status * 1,Refactor user processors code structure. 2,Remove user properties first name and last name. 3, Add new user properties of resource limit. 4, Changed related thrift structure from 'name' to 'id', 5,Add some comments for role. * To improve the drop space , on't delete useless role data * Rebased and added (raftex_obj raftex_thrift_obj wal_obj time_obj) to authentication_test module * Rebased and resolved some conflicts * Address laura-ding's comment * Resolved code alignment

#### What type of PR is this? - [ ] bug - [ ] feature - [x] enhancement #### What does this PR do? The original br tools require password-free ssh to each machine which add extra work for users. This pr's main goal is to remove ssh dependency. To do this, we introduce an agent in each machine. Then br can handle machines' data through agent's service. Agent communicate with the meta service through heartbeat. By heartbeat, agent register itself to meta service and pull the services it should supervise in its host. The agent: vesoft-inc/nebula-agent#1 The br: vesoft-inc/nebula-br#22 This pr includes: 1. refactor br related code, including renaming and adjust code structure. 2. batch the snapshot rpc by spaces 3. add agent heartbeat 4. report data/root path in the storaged/graphd heartbeat #### Which issue(s)/PR(s) this PR relates to? The agent: vesoft-inc/nebula-agent#1 The br: vesoft-inc/nebula-br#22 #### Special notes for your reviewer, ex. impact of this fix, etc: #### Additional context: #### Checklist： - [ ] Documentation affected （Please add the label if documentation needs to be modified.) - [ ] Incompatible （If it is incompatible, please describe it and add corresponding label.） - [ ] Need to cherry-pick （If need to cherry-pick to some branches, please label the destination version(s).) - [ ] Performance impacted: Consumes more CPU/Memory #### Release notes： Please confirm whether to reflect in release notes and how to describe: > ` Migrated from vesoft-inc#3469 Co-authored-by: pengwei.song <90180021+pengweisong@users.noreply.github.com>

boshengchen mentioned this pull request May 17, 2019

user management in Meta Server #257

Closed

darionyaphet reviewed May 17, 2019

View reviewed changes

src/meta/processors/AuthenticationProcessor.cpp Outdated Show resolved Hide resolved

darionyaphet reviewed May 17, 2019

View reviewed changes

src/interface/meta.thrift Outdated Show resolved Hide resolved

boshengchen force-pushed the user_meta branch 2 times, most recently from 6a91327 to 208dd1a Compare May 17, 2019 08:59

dangleptr suggested changes May 23, 2019

View reviewed changes

src/interface/meta.thrift Outdated Show resolved Hide resolved

src/interface/meta.thrift Show resolved Hide resolved

src/meta/MetaServiceUtils.cpp Outdated Show resolved Hide resolved

boshengchen force-pushed the user_meta branch from 851568b to 63db536 Compare May 27, 2019 06:40

boshengchen force-pushed the user_meta branch 3 times, most recently from 3a52188 to dbf79e2 Compare May 30, 2019 09:24

dangleptr reviewed May 30, 2019

View reviewed changes

This was referenced May 30, 2019

Support user manager in graph layer #480

Closed

Access rights control #190

Closed

dangleptr requested review from laura-ding, critical27 and dutor May 31, 2019 08:47

boshengchen force-pushed the user_meta branch from ec4e59d to dd89ad1 Compare June 3, 2019 07:20

boshengchen force-pushed the user_meta branch from dd89ad1 to 25813b6 Compare June 3, 2019 12:46

dangleptr reviewed Jun 4, 2019

View reviewed changes

dangleptr added this to the v1_beta_release milestone Jun 4, 2019

dangleptr requested a review from sherman-the-tank June 4, 2019 02:43

dangleptr previously approved these changes Jun 12, 2019

View reviewed changes

laura-ding reviewed Jun 14, 2019

View reviewed changes

src/meta/processors/usersMan/AuthenticationProcessor.cpp Outdated Show resolved Hide resolved

src/meta/test/CMakeLists.txt Outdated Show resolved Hide resolved

src/meta/test/TestUtils.h Show resolved Hide resolved

boshengchen dismissed dangleptr’s stale review via ee8875a June 14, 2019 09:02

boshengchen force-pushed the user_meta branch from 6215d7a to ee8875a Compare June 14, 2019 09:02

laura-ding reviewed Jun 14, 2019

View reviewed changes

src/meta/test/CMakeLists.txt Outdated Show resolved Hide resolved

boshengchen added 7 commits June 17, 2019 09:29

User manager of Meta Server

7e87c1e

1, Changed user data transport protocol from dataman to thrift. 2, Re…

0f22f7d

…moved user property email and phone. 3, Added new user property lock status

1,Refactor user processors code structure. 2,Remove user properties f…

27eec63

…irst name and last name. 3, Add new user properties of resource limit. 4, Changed related thrift structure from 'name' to 'id', 5,Add some comments for role.

To improve the drop space , on't delete useless role data

75f0ebe

Rebased and added (raftex_obj raftex_thrift_obj wal_obj time_obj) to …

dcf80c7

…authentication_test module

Rebased and resolved some conflicts

e561d8f

Address laura-ding's comment

2802590

boshengchen force-pushed the user_meta branch from ee8875a to c4eeebc Compare June 17, 2019 01:52

laura-ding reviewed Jun 19, 2019

View reviewed changes

Resolved code alignment

75052dd

boshengchen force-pushed the user_meta branch from c4eeebc to 75052dd Compare June 19, 2019 02:52

laura-ding approved these changes Jun 19, 2019

View reviewed changes

Merge branch 'master' into user_meta

baa5e6b

dangleptr approved these changes Jun 19, 2019

View reviewed changes

dangleptr merged commit 13a849e into vesoft-inc:master Jun 19, 2019

boshengchen deleted the user_meta branch June 19, 2019 09:40

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

User manager in meta #400

User manager in meta #400

boshengchen commented May 17, 2019

nebula-community-bot commented May 17, 2019

boshengchen commented May 24, 2019

boshengchen commented May 27, 2019

boshengchen commented May 31, 2019

dangleptr commented May 31, 2019

dangleptr commented May 31, 2019

boshengchen commented May 31, 2019 •

edited

Loading

boshengchen commented Jun 3, 2019

boshengchen commented Jun 3, 2019

dangleptr Jun 4, 2019

dangleptr Jun 4, 2019

boshengchen Jun 4, 2019

dangleptr Jun 4, 2019

dangleptr Jun 4, 2019

boshengchen Jun 4, 2019

dangleptr Jun 4, 2019

boshengchen Jun 4, 2019 •

edited

Loading

dangleptr Jun 4, 2019

boshengchen Jun 4, 2019

dangleptr Jun 4, 2019

boshengchen Jun 4, 2019

boshengchen commented Jun 14, 2019

dangleptr commented Jun 17, 2019

nebula-community-bot commented Jun 17, 2019

laura-ding Jun 19, 2019

boshengchen Jun 19, 2019

laura-ding left a comment

boshengchen commented Jun 19, 2019

dangleptr commented Jun 19, 2019

nebula-community-bot commented Jun 19, 2019

dangleptr left a comment

		}


		void AlterUserProcessor::process(const cpp2::AlterUserReq& req) {



		void ChangePasswordProcessor::process(const cpp2::ChangePasswordReq& req) {
		folly::SharedMutex::WriteHolder wHolder(LockUtils::userLock());

User manager in meta #400

User manager in meta #400

Conversation

boshengchen commented May 17, 2019

nebula-community-bot commented May 17, 2019

boshengchen commented May 24, 2019

boshengchen commented May 27, 2019

boshengchen commented May 31, 2019

dangleptr commented May 31, 2019

dangleptr commented May 31, 2019

boshengchen commented May 31, 2019 • edited Loading

boshengchen commented Jun 3, 2019

boshengchen commented Jun 3, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

boshengchen Jun 4, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

boshengchen commented Jun 14, 2019

dangleptr commented Jun 17, 2019

nebula-community-bot commented Jun 17, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

laura-ding left a comment

Choose a reason for hiding this comment

boshengchen commented Jun 19, 2019

dangleptr commented Jun 19, 2019

nebula-community-bot commented Jun 19, 2019

dangleptr left a comment

Choose a reason for hiding this comment

boshengchen commented May 31, 2019 •

edited

Loading

boshengchen Jun 4, 2019 •

edited

Loading