[Fix](Outfile) Fixed the problem that the concurrent Outfile wrote multiple Success files#33016
Conversation
|
Thank you for your contribution to Apache Doris. Since 2024-03-18, the Document has been moved to doris-website. |
|
clang-tidy review says "All clean, LGTM! 👍" |
|
run buildall |
1 similar comment
|
run buildall |
| } | ||
| } | ||
|
|
||
| void PInternalService::outfile_write_success(google::protobuf::RpcController* controller, |
There was a problem hiding this comment.
warning: method 'outfile_write_success' can be made static [readability-convert-member-functions-to-static]
| void PInternalService::outfile_write_success(google::protobuf::RpcController* controller, | |
| static void PInternalService::outfile_write_success(google::protobuf::RpcController* controller, |
TPC-H: Total hot run time: 38198 ms |
TPC-DS: Total hot run time: 182437 ms |
ClickBench: Total hot run time: 30.04 s |
|
Load test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G' |
|
TeamCity be ut coverage result: |
|
run buildall |
TPC-H: Total hot run time: 38972 ms |
TPC-DS: Total hot run time: 182186 ms |
ClickBench: Total hot run time: 31.25 s |
|
Load test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G' |
|
TeamCity be ut coverage result: |
|
run buildall |
TPC-H: Total hot run time: 38918 ms |
TPC-DS: Total hot run time: 183296 ms |
ClickBench: Total hot run time: 30.58 s |
|
Load test result on machine: 'aliyun_ecs.c7a.8xlarge_32C64G' |
|
TeamCity be ut coverage result: |
|
run buildall |
|
TeamCity be ut coverage result: |
|
PR approved by at least one committer and no changes requested. |
|
PR approved by anyone and no changes requested. |
…ltiple Success files (apache#33016) **Problem:** When we enable concurrent `Outfile` and specify the `success_file_name`, a SUCCESS file is written for each BE instance, which is not what we expected **solution:** Therefore, we added a new RPC request that when the Outfile was completed, the FE sent an RPC to the Be request to write one Success file.
…ltiple Success files (apache#33016) **Problem:** When we enable concurrent `Outfile` and specify the `success_file_name`, a SUCCESS file is written for each BE instance, which is not what we expected **solution:** Therefore, we added a new RPC request that when the Outfile was completed, the FE sent an RPC to the Be request to write one Success file.
…ltiple Success files (apache#33016) **Problem:** When we enable concurrent `Outfile` and specify the `success_file_name`, a SUCCESS file is written for each BE instance, which is not what we expected **solution:** Therefore, we added a new RPC request that when the Outfile was completed, the FE sent an RPC to the Be request to write one Success file.
…ltiple Success files (apache#33016) **Problem:** When we enable concurrent `Outfile` and specify the `success_file_name`, a SUCCESS file is written for each BE instance, which is not what we expected **solution:** Therefore, we added a new RPC request that when the Outfile was completed, the FE sent an RPC to the Be request to write one Success file.
…ltiple Success files apache#33016 (apache#33720)
) ### What problem does this PR solve? Related PR: #33016 Introduced from #33016, when specify the "schema" property in outfile clause with parquet format, it will return error: ``` Parquet schema number does not equal to select item number ``` This is because we wrongly analyze `OutfileClause` twice.
…che#57232) ### What problem does this PR solve? Related PR: apache#33016 Introduced from apache#33016, when specify the "schema" property in outfile clause with parquet format, it will return error: ``` Parquet schema number does not equal to select item number ``` This is because we wrongly analyze `OutfileClause` twice.
) ### What problem does this PR solve? Related PR: #33016 Introduced from #33016, when specify the "schema" property in outfile clause with parquet format, it will return error: ``` Parquet schema number does not equal to select item number ``` This is because we wrongly analyze `OutfileClause` twice.
…che#57232) Related PR: apache#33016 Introduced from apache#33016, when specify the "schema" property in outfile clause with parquet format, it will return error: ``` Parquet schema number does not equal to select item number ``` This is because we wrongly analyze `OutfileClause` twice.
Proposed changes
Issue Number: close #xxx
Problem:
When we enable concurrent
Outfileand specify thesuccess_file_name, a SUCCESS file is written for each BE instance, which is not what we expectedsolution:
Therefore, we added a new RPC request that when the Outfile was completed, the FE sent an RPC to the Be request to write one Success file.
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...