optimize: distinguish database behavior according to the branch type #2415

booogu · 2020-03-17T11:13:53Z

Ⅰ. Describe what this PR did

1.Change the interceptorType to branchType,optimize the logic in TccActionInterceptor
2.Optimize the judge condition of use UndoFunction in AT's ExecuteTemplate

Ⅱ. Does this pull request fix one issue?

Ⅲ. Why don't you add test cases (unit test/integration test)?

Ⅳ. Describe how to verify it

Ⅴ. Special notes for reviews

… is switched

codecov-io · 2020-03-17T11:22:35Z

Codecov Report

Merging #2415 into develop will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             develop    #2415   +/-   ##
==========================================
  Coverage      51.20%   51.20%           
  Complexity      2811     2811           
==========================================
  Files            554      554           
  Lines          17774    17774           
  Branches        2100     2100           
==========================================
  Hits            9102     9102           
  Misses          7812     7812           
  Partials         860      860

booogu · 2020-03-19T07:46:23Z

PTAL @zjinlei @slievrly

...c/main/java/io/seata/integration/dubbo/alibaba/AlibabaDubboTransactionPropagationFilter.java

.../dubbo/src/main/java/io/seata/integration/dubbo/ApacheDubboTransactionPropagationFilter.java

…f_tcc

.../dubbo/src/main/java/io/seata/integration/dubbo/ApacheDubboTransactionPropagationFilter.java

zjinlei · 2020-04-22T05:32:37Z

rm-datasource/src/main/java/io/seata/rm/datasource/exec/ExecuteTemplate.java

@@ -66,7 +68,7 @@
                                                     StatementCallback<T, S> statementCallback,
                                                     Object... args) throws SQLException {

-        if (!RootContext.inGlobalTransaction() && !RootContext.requireGlobalLock()) {
+        if (!requiresUndoFunction()) {


@GlobalLock does not have undo, this function name should be optimized

见：apache#2415

wangliang181230 · 2020-04-23T07:14:47Z

...c/main/java/io/seata/integration/dubbo/alibaba/AlibabaDubboTransactionPropagationFilter.java

        } else {
            if (rpcXid != null) {
                RootContext.bind(rpcXid);
-                RootContext.bindInterceptorType(rpcXidInterceptorType);
+                if (StringUtils.equals(BranchType.TCC.name(), rpcBranchType)) {
+                    RootContext.bindBranchType(BranchType.TCC);


Saga branch never set ?

I have discussed with the author of saga. This pr is only responsible for the branchType judgment logic in ExecuteTemplate and the TCC branchType passing in Filter. The specific Saga branchtype assignment and other logic will be provided by the feature saga related pr

this #2551, autor is wangliang1986

long187

LGTM

…f_tcc

wangliang181230

LGTM

zjinlei

LGTM

slievrly · 2020-04-26T10:19:15Z

...c/main/java/io/seata/integration/dubbo/alibaba/AlibabaDubboTransactionPropagationFilter.java

        if (LOGGER.isDebugEnabled()) {
            LOGGER.debug("xid in RootContext[{}] xid in RpcContext[{}]", xid, rpcXid);
        }
        boolean bind = false;
        if (xid != null) {
            RpcContext.getContext().setAttachment(RootContext.KEY_XID, xid);
-            RpcContext.getContext().setAttachment(RootContext.KEY_XID_INTERCEPTOR_TYPE, xidInterceptorType);
+            RpcContext.getContext().setAttachment(RootContext.KEY_BRANCH_TYPE, branchType);


Is it better to separate and judge?

It will not be processed this time, and transaction context object encapsulation will be provided later to handle log output uniformly.

slievrly · 2020-04-26T10:20:56Z

...c/main/java/io/seata/integration/dubbo/alibaba/AlibabaDubboTransactionPropagationFilter.java

                if (LOGGER.isDebugEnabled()) {
-                    LOGGER.debug("unbind[{}] interceptorType[{}] from RootContext", unbindXid, unbindInterceptorType);
+                    LOGGER.debug("unbind xid [{}] branchType [{}] from RootContext", unbindXid, previousBranchType);


BranchType print null in AT mode is not very elegant.

How about judge in the log output temporarily,and print "AT" if it is null

slievrly · 2020-04-26T10:21:53Z

rm-datasource/src/main/java/io/seata/rm/datasource/exec/ExecuteTemplate.java

@@ -115,4 +117,17 @@
        }
        return rs;
    }
+
+    private static boolean requiresLockOrUndoFunction() {


Other more meaningful method name?

Change it to "shouldExecuteInATMode"

…f_tcc

slievrly

LGTM

wangliang181230 · 2020-04-29T03:25:07Z

此PR合并后，出现的BUG: 在TCC或SAGA事务内，还执行了SQL识别功能，导致SAGA事务执行失败。
代码行：AbstractConnectionProxy类的PreparedStatement prepareStatement(String sql)方法，第112行。
@Override public PreparedStatement prepareStatement(String sql) throws SQLException { String dbType = getDbType(); // support oracle 10.2+ PreparedStatement targetPreparedStatement = null; if (RootContext.inGlobalTransaction()) { List<SQLRecognizer> sqlRecognizers = SQLVisitorFactory.get(sql, dbType); if (sqlRecognizers != null && sqlRecognizers.size() == 1) { SQLRecognizer sqlRecognizer = sqlRecognizers.get(0); if (sqlRecognizer != null && sqlRecognizer.getSQLType() == SQLType.INSERT) { String tableName = ColumnUtils.delEscape(sqlRecognizer.getTableName(), dbType); TableMeta tableMeta = TableMetaCacheFactory.getTableMetaCache(dbType).getTableMeta(getTargetConnection(), tableName, getDataSourceProxy().getResourceId()); targetPreparedStatement = getTargetConnection().prepareStatement(sql, new String[]{tableMeta.getPkName()}); } } } if (targetPreparedStatement == null) { targetPreparedStatement = getTargetConnection().prepareStatement(sql); } return new PreparedStatementProxy(this, targetPreparedStatement, sql); }

中的RootContext.inGlobalTransaction()是否应该改为ExecuteTemplate.shouldExecuteInATMode()。

booogu · 2020-04-29T03:34:20Z

此PR合并后，出现的BUG: 在TCC或SAGA事务内，还执行了SQL识别功能，导致SAGA事务执行失败。
代码行：AbstractConnectionProxy类的PreparedStatement prepareStatement(String sql)方法，第112行。
@Override public PreparedStatement prepareStatement(String sql) throws SQLException { String dbType = getDbType(); // support oracle 10.2+ PreparedStatement targetPreparedStatement = null; if (RootContext.inGlobalTransaction()) { List<SQLRecognizer> sqlRecognizers = SQLVisitorFactory.get(sql, dbType); if (sqlRecognizers != null && sqlRecognizers.size() == 1) { SQLRecognizer sqlRecognizer = sqlRecognizers.get(0); if (sqlRecognizer != null && sqlRecognizer.getSQLType() == SQLType.INSERT) { String tableName = ColumnUtils.delEscape(sqlRecognizer.getTableName(), dbType); TableMeta tableMeta = TableMetaCacheFactory.getTableMetaCache(dbType).getTableMeta(getTargetConnection(), tableName, getDataSourceProxy().getResourceId()); targetPreparedStatement = getTargetConnection().prepareStatement(sql, new String[]{tableMeta.getPkName()}); } } } if (targetPreparedStatement == null) { targetPreparedStatement = getTargetConnection().prepareStatement(sql); } return new PreparedStatementProxy(this, targetPreparedStatement, sql); }

中的RootContext.inGlobalTransaction()是否应该改为ExecuteTemplate.shouldExecuteInATMode()。

抱歉给您带来了问题，是的，确实有这个问题，虽然执行的时候不进行sql解析构建undoLog了，但是这里还是会执行sql识别，这是我的疏忽，我马上提交一个修复pr

wangliang181230 · 2020-04-29T03:59:53Z

OK

…pache#2415)

booogu · 2020-05-02T07:04:41Z

OK
I fix it in this pr,please take a look.
#2650

booogu and others added 14 commits February 26, 2020 23:23

optimize: optimize transaction context behavior when transaction mode…

92a1e4b

… is switched

adjust BranchType switching with Propagation feature

7829ce7

Merge remote-tracking branch 'upstream/develop' into f_tcc

f4e4078

format code style

ef1ea3c

Merge branch 'develop' into f_tcc

65312dd

Merge branch 'develop' into f_tcc

d3f1483

fix the mistake in sagaBranch and rename defaultBranchType

9bbdd1f

handle required branchType

8fb6dc2

Merge branch 'develop' into f_tcc

0596108

Merge branch 'develop' into f_tcc

52335c8

Merge branch 'develop' into f_tcc

57e7dfd

use StringUtils

06e79ea

Merge remote-tracking branch 'upstream/develop' into f_tcc

d0b3128

optimize: optimize interceptorType to branchType

e721fb1

booogu added 2 commits March 17, 2020 19:57

remove unuse

f6a1cb6

optimize rootcontext

03d9a95

zjinlei added this to the 1.2.0 milestone Mar 18, 2020

booogu added 6 commits March 19, 2020 11:01

optimize branchType to tccScope

a3837de

Merge remote-tracking branch 'upstream/develop' into f_tcc

6030169

optimize branchType and tccScope

fd12c62

format some incorrect code

c50587f

format code style

97510c0

format code style

657c45e

zjinlei and others added 4 commits March 19, 2020 18:15

Merge branch 'develop' into f_tcc

3d9a032

Merge branch 'develop' into f_tcc

5afdfc2

Merge remote-tracking branch 'upstream/develop' into f_tcc

d5dd5cd

merge latest code

3a0613b

zjinlei removed this from the 1.2.0 milestone Apr 10, 2020

zjinlei reviewed Apr 21, 2020

View reviewed changes

...c/main/java/io/seata/integration/dubbo/alibaba/AlibabaDubboTransactionPropagationFilter.java Outdated Show resolved Hide resolved

zjinlei reviewed Apr 21, 2020

View reviewed changes

.../dubbo/src/main/java/io/seata/integration/dubbo/ApacheDubboTransactionPropagationFilter.java Outdated Show resolved Hide resolved

booogu added 2 commits April 22, 2020 10:22

format code and keep code same in integration module

9979b00

Merge branch 'f_tcc' of https://github.com/CharmingRabbit/seata into …

95ddbe2

…f_tcc

zjinlei reviewed Apr 22, 2020

View reviewed changes

wangliang181230 added a commit to wangliang181230/seata that referenced this pull request Apr 22, 2020

AT事务与其他部分事务隔离的代码已在另一个PR中实现。现在恢复部分代码。

9188326

见：apache#2415

wangliang181230 reviewed Apr 23, 2020

View reviewed changes

long187 approved these changes Apr 23, 2020

View reviewed changes

zjinlei and others added 3 commits April 24, 2020 21:14

Merge branch 'develop' into f_tcc

7666bc1

optimize log output and func name

127b3e8

Merge branch 'f_tcc' of https://github.com/CharmingRabbit/seata into …

d7d249b

…f_tcc

booogu changed the title ~~optimize: optimize interceptorType to branchType~~ optimize: distinguish database behavior according to the branch type Apr 24, 2020

wangliang181230 approved these changes Apr 24, 2020

View reviewed changes

restore

69f29b7

zjinlei approved these changes Apr 25, 2020

View reviewed changes

zjinlei added the Do Not Merge Do not merge into develop label Apr 25, 2020

Merge branch 'develop' into f_tcc

7e86dd3

slievrly reviewed Apr 26, 2020

View reviewed changes

booogu added 2 commits April 26, 2020 19:07

optimize func name and log output

f5c438e

Merge branch 'f_tcc' of https://github.com/CharmingRabbit/seata into …

7d41df6

…f_tcc

slievrly approved these changes Apr 26, 2020

View reviewed changes

remove travis -p param

970e86f

zjinlei merged commit e5c68a7 into apache:develop Apr 27, 2020

xiaowan pushed a commit to xiaowan/seata that referenced this pull request Apr 30, 2020

optimize: distinguish database behavior according to the branch type (a…

d88b9d3

…pache#2415)

hdr893487191 mentioned this pull request Jun 3, 2020

seata 是不是要支持 spring cloud 使用tcc模式？ #2754

Closed

booogu deleted the f_tcc branch February 2, 2021 01:58

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

optimize: distinguish database behavior according to the branch type #2415

optimize: distinguish database behavior according to the branch type #2415

booogu commented Mar 17, 2020

codecov-io commented Mar 17, 2020 •

edited

booogu commented Mar 19, 2020

zjinlei Apr 22, 2020

wangliang181230 Apr 23, 2020

booogu Apr 23, 2020

long187 Apr 23, 2020

long187 left a comment

wangliang181230 left a comment

zjinlei left a comment

slievrly Apr 26, 2020

booogu Apr 26, 2020

slievrly Apr 26, 2020

booogu Apr 26, 2020

slievrly Apr 26, 2020

booogu Apr 26, 2020

slievrly left a comment

wangliang181230 commented Apr 29, 2020

booogu commented Apr 29, 2020

wangliang181230 commented Apr 29, 2020

booogu commented May 2, 2020

optimize: distinguish database behavior according to the branch type #2415

optimize: distinguish database behavior according to the branch type #2415

Conversation

booogu commented Mar 17, 2020

Ⅰ. Describe what this PR did

Ⅱ. Does this pull request fix one issue?

Ⅲ. Why don't you add test cases (unit test/integration test)?

Ⅳ. Describe how to verify it

Ⅴ. Special notes for reviews

codecov-io commented Mar 17, 2020 • edited

Codecov Report

booogu commented Mar 19, 2020

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

long187 left a comment

Choose a reason for hiding this comment

wangliang181230 left a comment

Choose a reason for hiding this comment

zjinlei left a 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

Choose a reason for hiding this comment

slievrly left a comment

Choose a reason for hiding this comment

wangliang181230 commented Apr 29, 2020

booogu commented Apr 29, 2020

wangliang181230 commented Apr 29, 2020

booogu commented May 2, 2020

codecov-io commented Mar 17, 2020 •

edited