You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@doris.apache.org by GitBox <gi...@apache.org> on 2021/11/29 05:13:16 UTC
[GitHub] [incubator-doris] caiconghui opened a new pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
caiconghui opened a new pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240
…
## Proposed changes
Describe the big picture of your changes here to communicate to the maintainers why we should accept this pull request. If it fixes a bug or resolves a feature request, be sure to link to that issue.
## Types of changes
What types of changes does your code introduce to Doris?
_Put an `x` in the boxes that apply_
- [ ] Bugfix (non-breaking change which fixes an issue)
- [ ] New feature (non-breaking change which adds functionality)
- [ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
- [ ] Documentation Update (if none of the other choices apply)
- [ ] Code refactor (Modify the code structure, format the code, etc...)
- [ ] Optimization. Including functional usability improvements and performance improvements.
- [ ] Dependency. Such as changes related to third-party components.
- [ ] Other.
## Checklist
_Put an `x` in the boxes that apply. You can also fill these out after creating the PR. If you're unsure about any of them, don't hesitate to ask. We're here to help! This is simply a reminder of what we are going to look for before merging your code._
- [x] I have created an issue on (Fix #7239 ) and described the bug/feature there in detail
- [ ] Compiling and unit tests pass locally with my changes
- [ ] I have added tests that prove my fix is effective or that my feature works
- [ ] If these changes need document changes, I have updated the document
- [ ] Any dependent changes have been merged
## Further comments
If this is a relatively large or complex change, kick off the discussion at dev@doris.apache.org by explaining why you chose the solution you did and what alternatives you considered, etc...
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] github-actions[bot] commented on pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
github-actions[bot] commented on pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240#issuecomment-1004616175
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] morningman merged pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
morningman merged pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] caiconghui commented on a change in pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
caiconghui commented on a change in pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240#discussion_r777792407
##########
File path: fe/fe-core/src/main/java/org/apache/doris/load/update/UpdateStmtExecutor.java
##########
@@ -177,21 +176,7 @@ private void executePlan() throws Exception {
private void commitAndPublishTxn() throws UserException {
GlobalTransactionMgr globalTransactionMgr = Catalog.getCurrentGlobalTransactionMgr();
- // situation1: no data is updated, abort transaction
- if (effectRows == 0) {
- LOG.info("abort transaction for update stmt, query id:{}, reason: {}", DebugUtil.printId(queryId),
- TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- globalTransactionMgr.abortTransaction(dbId, txnId, TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- StringBuilder sb = new StringBuilder();
- sb.append("{'label':'").append(label);
- sb.append(", 'txnId':'").append(txnId).append("'");
- sb.append(", 'queryId':'").append(DebugUtil.printId(queryId)).append("'");
- sb.append("}");
- analyzer.getContext().getState().setOk(effectRows, 0, sb.toString());
- return;
Review comment:
@yiguolei now, commit and publish txn without any tablet almost cost nothing, we should unity it to avoid user to load data again, because no partition with data load may have some cases for normal and unnormal situation, and we may write extra code to handle this case for insert and update, which make code look like ugly and may not easy to handle
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] yiguolei commented on a change in pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
yiguolei commented on a change in pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240#discussion_r777794248
##########
File path: fe/fe-core/src/main/java/org/apache/doris/load/update/UpdateStmtExecutor.java
##########
@@ -177,21 +176,7 @@ private void executePlan() throws Exception {
private void commitAndPublishTxn() throws UserException {
GlobalTransactionMgr globalTransactionMgr = Catalog.getCurrentGlobalTransactionMgr();
- // situation1: no data is updated, abort transaction
- if (effectRows == 0) {
- LOG.info("abort transaction for update stmt, query id:{}, reason: {}", DebugUtil.printId(queryId),
- TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- globalTransactionMgr.abortTransaction(dbId, txnId, TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- StringBuilder sb = new StringBuilder();
- sb.append("{'label':'").append(label);
- sb.append(", 'txnId':'").append(txnId).append("'");
- sb.append(", 'queryId':'").append(DebugUtil.printId(queryId)).append("'");
- sb.append("}");
- analyzer.getContext().getState().setOk(effectRows, 0, sb.toString());
- return;
Review comment:
Yes, I agree with you.
Then LGTM.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] caiconghui commented on a change in pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
caiconghui commented on a change in pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240#discussion_r777280205
##########
File path: fe/fe-core/src/main/java/org/apache/doris/load/update/UpdateStmtExecutor.java
##########
@@ -177,21 +176,7 @@ private void executePlan() throws Exception {
private void commitAndPublishTxn() throws UserException {
GlobalTransactionMgr globalTransactionMgr = Catalog.getCurrentGlobalTransactionMgr();
- // situation1: no data is updated, abort transaction
- if (effectRows == 0) {
- LOG.info("abort transaction for update stmt, query id:{}, reason: {}", DebugUtil.printId(queryId),
- TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- globalTransactionMgr.abortTransaction(dbId, txnId, TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- StringBuilder sb = new StringBuilder();
- sb.append("{'label':'").append(label);
- sb.append(", 'txnId':'").append(txnId).append("'");
- sb.append(", 'queryId':'").append(DebugUtil.printId(queryId)).append("'");
- sb.append("}");
- analyzer.getContext().getState().setOk(effectRows, 0, sb.toString());
- return;
Review comment:
yes, i have tested stream load and insert already, it works well. actually, we should keep same behavior with mysql, whether update or insert 0 rows
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] caiconghui commented on a change in pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
caiconghui commented on a change in pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240#discussion_r777279965
##########
File path: fe/fe-core/src/main/java/org/apache/doris/load/routineload/KafkaRoutineLoadJob.java
##########
@@ -240,16 +240,6 @@ protected boolean checkCommitInfo(RLTaskTxnCommitAttachment rlTaskTxnCommitAttac
return true;
}
- if (txnStatusChangeReason != null && txnStatusChangeReason == TransactionState.TxnStatusChangeReason.NO_PARTITIONS) {
- // Because the max_filter_ratio of routine load task is always 1.
- // Therefore, under normal circumstances, routine load task will not return the error "too many filtered rows".
- // If no data is imported, the error "all partitions have no load data" may only be returned.
- // In this case, the status of the transaction is ABORTED,
- // but we still need to update the offset to skip these error lines.
- Preconditions.checkState(txnState.getTransactionStatus() == TransactionStatus.ABORTED, txnState.getTransactionStatus());
- return true;
Review comment:
when we can commit empty rowset, txnStatusChangeReason would be null
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] morningman commented on a change in pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
morningman commented on a change in pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240#discussion_r777278929
##########
File path: fe/fe-core/src/main/java/org/apache/doris/load/update/UpdateStmtExecutor.java
##########
@@ -177,21 +176,7 @@ private void executePlan() throws Exception {
private void commitAndPublishTxn() throws UserException {
GlobalTransactionMgr globalTransactionMgr = Catalog.getCurrentGlobalTransactionMgr();
- // situation1: no data is updated, abort transaction
- if (effectRows == 0) {
- LOG.info("abort transaction for update stmt, query id:{}, reason: {}", DebugUtil.printId(queryId),
- TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- globalTransactionMgr.abortTransaction(dbId, txnId, TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- StringBuilder sb = new StringBuilder();
- sb.append("{'label':'").append(label);
- sb.append(", 'txnId':'").append(txnId).append("'");
- sb.append(", 'queryId':'").append(DebugUtil.printId(queryId)).append("'");
- sb.append("}");
- analyzer.getContext().getState().setOk(effectRows, 0, sb.toString());
- return;
Review comment:
I remember that Doris do not support commit empty rowset. Have you tested this case yet?
##########
File path: fe/fe-core/src/main/java/org/apache/doris/load/routineload/KafkaRoutineLoadJob.java
##########
@@ -240,16 +240,6 @@ protected boolean checkCommitInfo(RLTaskTxnCommitAttachment rlTaskTxnCommitAttac
return true;
}
- if (txnStatusChangeReason != null && txnStatusChangeReason == TransactionState.TxnStatusChangeReason.NO_PARTITIONS) {
- // Because the max_filter_ratio of routine load task is always 1.
- // Therefore, under normal circumstances, routine load task will not return the error "too many filtered rows".
- // If no data is imported, the error "all partitions have no load data" may only be returned.
- // In this case, the status of the transaction is ABORTED,
- // but we still need to update the offset to skip these error lines.
- Preconditions.checkState(txnState.getTransactionStatus() == TransactionStatus.ABORTED, txnState.getTransactionStatus());
- return true;
Review comment:
We need to return true to skip those unloaded rows.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org
[GitHub] [incubator-doris] yiguolei commented on a change in pull request #7240: [Load] Transaction for load job with no data for all partitions should be considered as normal and should not be aborted
Posted by GitBox <gi...@apache.org>.
yiguolei commented on a change in pull request #7240:
URL: https://github.com/apache/incubator-doris/pull/7240#discussion_r777787114
##########
File path: fe/fe-core/src/main/java/org/apache/doris/load/update/UpdateStmtExecutor.java
##########
@@ -177,21 +176,7 @@ private void executePlan() throws Exception {
private void commitAndPublishTxn() throws UserException {
GlobalTransactionMgr globalTransactionMgr = Catalog.getCurrentGlobalTransactionMgr();
- // situation1: no data is updated, abort transaction
- if (effectRows == 0) {
- LOG.info("abort transaction for update stmt, query id:{}, reason: {}", DebugUtil.printId(queryId),
- TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- globalTransactionMgr.abortTransaction(dbId, txnId, TransactionCommitFailedException.NO_DATA_TO_LOAD_MSG);
- StringBuilder sb = new StringBuilder();
- sb.append("{'label':'").append(label);
- sb.append(", 'txnId':'").append(txnId).append("'");
- sb.append(", 'queryId':'").append(DebugUtil.printId(queryId)).append("'");
- sb.append("}");
- analyzer.getContext().getState().setOk(effectRows, 0, sb.toString());
- return;
Review comment:
Yes， Doris support empty rowset. Because when there are 10 tablets, but user could insert 1 row, and this row only exist in one tablet and other tablets have empty rowset.
But, we should optimize this. If all tablets have empty rowsets, then just return success, not commit it, and abort the txn.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org