[Fix](partial update) Persist partial_update_info in RocksDB in case of BE restart after a partial update has commited #38331
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
Thank you for your contribution to Apache Doris. Since 2024-03-18, the Document has been moved to doris-website. |
bobhan1
force-pushed
the
persist-partial-update-info
branch
from
236a525 to
e79f107
Compare
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
|
|
||
| #include "olap/partial_update_info.h" | ||
|
|
||
| #include <gen_cpp/olap_file.pb.h> |
There was a problem hiding this comment.
warning: 'gen_cpp/olap_file.pb.h' file not found [clang-diagnostic-error]
#include <gen_cpp/olap_file.pb.h>
^|
|
||
| struct PartialUpdateInfo { | ||
| void init(const TabletSchema& tablet_schema, bool partial_update, | ||
| const std::set<string>& partial_update_cols, bool is_strict_mode, | ||
| const std::set<std::string>& partial_update_cols, bool is_strict_mode, |
There was a problem hiding this comment.
warning: no template named 'set' in namespace 'std' [clang-diagnostic-error]
const std::set<std::string>& partial_update_cols, bool is_strict_mode,
^| @@ -18,6 +18,8 @@ | |||
| #ifndef DORIS_BE_SRC_OLAP_ROWSET_ROWSET_META_MANAGER_H | |||
| #define DORIS_BE_SRC_OLAP_ROWSET_ROWSET_META_MANAGER_H | |||
|
|
|||
| #include <gen_cpp/olap_file.pb.h> | |||
There was a problem hiding this comment.
warning: 'gen_cpp/olap_file.pb.h' file not found [clang-diagnostic-error]
#include <gen_cpp/olap_file.pb.h>
^
|
run buildall |
1 similar comment
|
run buildall |
bobhan1
force-pushed
the
persist-partial-update-info
branch
from
ad2d9a9 to
efe2ea6
Compare
|
run buildall |
bobhan1
force-pushed
the
persist-partial-update-info
branch
from
efe2ea6 to
762e3ef
Compare
|
run buildall |
zhannngchen
reviewed
Jul 30, 2024
bobhan1
force-pushed
the
persist-partial-update-info
branch
2 times, most recently
from
68d404b to
4095139
Compare
|
run buildall |
bobhan1
force-pushed
the
persist-partial-update-info
branch
from
4095139 to
f9af921
Compare
bobhan1
force-pushed
the
persist-partial-update-info
branch
from
f9af921 to
d4b8f3b
Compare
github-actions
bot
added
the
approved
|
PR approved by at least one committer and no changes requested. |
|
PR approved by anyone and no changes requested. |
github-actions
bot
removed
the
approved
bobhan1
force-pushed
the
persist-partial-update-info
branch
from
be9266f to
628fb44
Compare
|
run buildall |
github-actions
bot
added
the
approved
github-actions
bot
added
the
approved
|
PR approved by at least one committer and no changes requested. |
Yukang-Lian
approved these changes
Aug 5, 2024
This was referenced Aug 7, 2024
zhannngchen
pushed a commit
that referenced
this pull request
Aug 8, 2024
dataroaring
pushed a commit
that referenced
this pull request
Aug 9, 2024
dataroaring
pushed a commit
that referenced
this pull request
Aug 11, 2024
…of BE restart after a partial update has commited (#38331) ## Proposed changes If a partial update has conflict with another load during publish phase, it should combine the two load's data into one to get the corrrect result. This procedure needs partial update info. But If BE crashed after the partial update load has committed, the partial update info will be missing becasuse it's not persisted and will not be restored in `DataDir::load()`. This PR persists partial update info in RocksDB before the txn is commited and remove it after the publish phase. Before #25147, partial update info is persisted with tablet_schema in RocksDB. #25147 split partial update info from tablet schema but forget to handle the persistence logic.
wyxxxcat
pushed a commit
to wyxxxcat/doris
that referenced
this pull request
Aug 14, 2024
…of BE restart after a partial update has commited (apache#38331) ## Proposed changes If a partial update has conflict with another load during publish phase, it should combine the two load's data into one to get the corrrect result. This procedure needs partial update info. But If BE crashed after the partial update load has committed, the partial update info will be missing becasuse it's not persisted and will not be restored in `DataDir::load()`. This PR persists partial update info in RocksDB before the txn is commited and remove it after the publish phase. Before apache#25147, partial update info is persisted with tablet_schema in RocksDB. apache#25147 split partial update info from tablet schema but forget to handle the persistence logic.
wyxxxcat
pushed a commit
to wyxxxcat/doris
that referenced
this pull request
Aug 14, 2024
…inter (apache#39066) fix member call on null pointer in apache#38331
dataroaring
pushed a commit
that referenced
this pull request
Aug 16, 2024
…of BE restart after a partial update has commited (#38331) ## Proposed changes If a partial update has conflict with another load during publish phase, it should combine the two load's data into one to get the corrrect result. This procedure needs partial update info. But If BE crashed after the partial update load has committed, the partial update info will be missing becasuse it's not persisted and will not be restored in `DataDir::load()`. This PR persists partial update info in RocksDB before the txn is commited and remove it after the publish phase. Before #25147, partial update info is persisted with tablet_schema in RocksDB. #25147 split partial update info from tablet schema but forget to handle the persistence logic.
GoGoWen
pushed a commit
to GoGoWen/incubator-doris
that referenced
this pull request
Aug 27, 2024
… in RocksDB in case of BE restart after a partial update has commited apache#38331" (apache#39078) picks apache#38331 and apache#39066
bobhan1
added a commit
to bobhan1/doris
that referenced
this pull request
Jan 21, 2025
… in RocksDB in case of BE restart after a partial update has commited apache#38331" (apache#39078) picks apache#38331 and apache#39066
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Proposed changes
If a partial update has conflict with another load during publish phase, it should combine the two load's data into one to get the corrrect result. This procedure needs partial update info. But If BE crashed after the partial update load has committed, the partial update info will be missing becasuse it's not persisted and will not be restored in
DataDir::load(). This PR persists partial update info in RocksDB before the txn is commited and remove it after the publish phase.Before #25147, partial update info is persisted with tablet_schema in RocksDB. #25147 split partial update info from tablet schema but forget to handle the persistence logic.
branch-2.1-pick: #39035
branch-2.0-pick: #39078