HBASE-26271: Cleanup the broken store files under data directory #3786
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
|
@Apache9, @wchevreuil This is the commit only containing changes related to the CleanerChore and exposing what is absolutely necessary for it's functionality. |
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
Apache9
reviewed
Oct 24, 2021
| * This Chore, every time it runs, will clear the unsused HFiles in the data | ||
| * folder. | ||
| */ | ||
| @InterfaceAudience.Private public class FileBasedStoreFileCleaner extends ScheduledChore { |
There was a problem hiding this comment.
Let's not name it 'FileBased'. I think it should be more general, maybe 'BrokenStoreFileCleaner'.
For now, the only condition is whether we will write to the data directory directly. No matter what is the actual store file tracker implementation, if it writes to the data directory directly, then we need this cleaner, otherwise we do not.
There was a problem hiding this comment.
Ok, I'm going with BrokenStoreFileCleaner then.
There was a problem hiding this comment.
Just another name suggestion: "LeftoversStoreFileCleaner"... Up to you, @BukrosSzabolcs !
| for (HRegion region : regionServer.getRegions()) { | ||
| for (HStore store : region.getStores()) { | ||
| //only clean do cleanup in store using file based storefile tracking | ||
| if (store.getStoreEngine().requireWritingToTmpDirFirst()) { |
There was a problem hiding this comment.
OK, good, this is exactly what I expected. So the only problem is the naming. Let's find a better name.
| deleteFile(file, store, deletedFiles, failedDeletes); | ||
| } | ||
|
|
||
| private boolean isCompactingFile(FileStatus file, HStore store) { |
There was a problem hiding this comment.
I suppose we should move the below isXXXFile to HRegion or HStore, with the protection of some locks. Otherwise we may have race, or at least, we depend on a very flaky order of the testing of each condition.
And do we really need to test isCompactingFile here? The compactedFile will not be tracked but compacting files should always be tracked?
There was a problem hiding this comment.
I suppose we should move the below isXXXFile to HRegion or HStore, with the protection of some locks. Otherwise we may have race, or at least, we depend on a very flaky order of the testing of each condition.
I guess the isOldEnough check would prevent that problem, as we would only care about files older than DEFAULT_FILEBASED_STOREFILE_CLEANER_TTL?
And do we really need to test isCompactingFile here? The compactedFile will not be tracked but compacting files should always be tracked?
Yeah, this seems redundant with the isActiveStorefile check.
There was a problem hiding this comment.
And do we really need to test isCompactingFile here? The compactedFile will not be tracked but compacting files should always be tracked?
The confusion might come from suboptimal naming.
isActiveStorefile -> lists the currently active storefiles
isCompactedFile -> are hfiles that got compacted, no longer active storefiles, but were not deleted yet (deletion is handled by a separate subsystem, so we should not touch them)
isCompactingFile -> file(s) a currently running compaction is writing into. This will become the new storefile when the compaction is done. It is checked to make sure we do not break stuck/longrunning compactions even is they are stuck/idle for more than the configured TTL.
I think we do need all of these checks.
I suppose we should move the below isXXXFile to HRegion or HStore, with the protection of some locks. Otherwise we may have race, or at least, we depend on a very flaky order of the testing of each condition.
As Wellington mentioned, the TTL should make sure we filter out any currently handled hfile. Any file we check afterwards was not touched in 12 hours. I think the only usecase where this might be an issue is a stuck compaction where we check isCompactingFile, but that method (getCompactionTargets) is synchronised so should be safe. Do you see any other possible issue?
There was a problem hiding this comment.
The confusion might come from suboptimal naming.
isActiveStorefile -> lists the currently active storefiles
isCompactedFile -> are hfiles that got compacted, no longer active storefiles, but were not deleted yet (deletion is handled by a separate subsystem, so we should not touch them)
isCompactingFile -> file(s) a currently running compaction is writing into. This will become the new storefile when the compaction is done. It is checked to make sure we do not break stuck/longrunning compactions even is they are stuck/idle for more than the configured TTL.
I think we do need all of these checks.
Thanks for clarifying. Thinking again, maybe it's worth having this extra check, for the exceptional cases compaction can last longer than the time threshold.
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/Compactor.java
Show resolved
Hide resolved
...er/src/main/java/org/apache/hadoop/hbase/regionserver/storefiletracker/StoreFileTracker.java
Outdated
Show resolved
Hide resolved
...rc/main/java/org/apache/hadoop/hbase/regionserver/storefiletracker/StoreFileTrackerBase.java
Outdated
Show resolved
Hide resolved
wchevreuil
reviewed
Oct 25, 2021
| deleteFile(file, store, deletedFiles, failedDeletes); | ||
| } | ||
|
|
||
| private boolean isCompactingFile(FileStatus file, HStore store) { |
There was a problem hiding this comment.
I suppose we should move the below isXXXFile to HRegion or HStore, with the protection of some locks. Otherwise we may have race, or at least, we depend on a very flaky order of the testing of each condition.
I guess the isOldEnough check would prevent that problem, as we would only care about files older than DEFAULT_FILEBASED_STOREFILE_CLEANER_TTL?
And do we really need to test isCompactingFile here? The compactedFile will not be tracked but compacting files should always be tracked?
Yeah, this seems redundant with the isActiveStorefile check.
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java
Outdated
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/Compactor.java
Outdated
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/Compactor.java
Show resolved
Hide resolved
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
|
The failing tests seem to be flaky. I was able to successfully run them locally. |
joshelser
requested changes
Nov 3, 2021
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/AbstractMultiFileWriter.java
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/BrokenStoreFileCleaner.java
Outdated
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/BrokenStoreFileCleaner.java
Outdated
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/BrokenStoreFileCleaner.java
Outdated
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/StoreEngine.java
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/Compactor.java
Outdated
Show resolved
Hide resolved
...-server/src/main/java/org/apache/hadoop/hbase/regionserver/compactions/DefaultCompactor.java
Outdated
Show resolved
Hide resolved
hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/TestBrokenStoreFileCleaner.java
Outdated
Show resolved
Hide resolved
hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/TestBrokenStoreFileCleaner.java
Show resolved
Hide resolved
hbase-server/src/test/java/org/apache/hadoop/hbase/regionserver/TestBrokenStoreFileCleaner.java
Outdated
Show resolved
Hide resolved
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
Just to make my intent clear, I think with a test method to exercise the TTL effectiveness and the minor code-formatting cleanup, this is fine to commit. Not sure if @Apache9 has more opinions on this change before he'd like to see it merged to the feature branch. |
|
💔 -1 overall
This message was automatically generated. |
joshelser
approved these changes
Nov 4, 2021
Apache9
reviewed
Nov 5, 2021
| return this.enabled.get(); | ||
| } | ||
|
|
||
| @InterfaceAudience.Private |
There was a problem hiding this comment.
Do we need this extra IA annotation? The class itself is IA.Private.
There was a problem hiding this comment.
You are right, deleting it.
| return; | ||
| } | ||
|
|
||
| if(isCompactingFile(file, store)){ |
There was a problem hiding this comment.
The 'compacting file' has its meaning in HBase. There is a filesCompacting field in HStore class, which is the files being compacted currently, not the files written out. So here let's change the name, otherwise it will confuse people.
There was a problem hiding this comment.
I'm renaming it to "isCompactionResultFile"
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/BrokenStoreFileCleaner.java
Show resolved
Hide resolved
...main/java/org/apache/hadoop/hbase/regionserver/compactions/AbstractMultiOutputCompactor.java
Show resolved
Hide resolved
|
#3700 has been merged and I rebased HBASE-26067 with the newest master. Please rebase the PR too. Thanks for your patient. We are very close to merge this PR. |
… tracker (apache#3681) Signed-off-by: Josh Elser <elserj@apache.org>
Add new chore to delete lefotver files in case file based storefile handling is used Expose the target files of currently running compactions for easier validation
fixes based on feedback
added javadoc fixed typos additonal fileTTL UT
improve naming and add some explanation in the comments
clean up after rebase
BukrosSzabolcs
force-pushed
the
HBASE-26271-2
branch
from
a86b9ca to
ae0fa65
Compare
|
@Apache9 Rebased my changes |
|
🎊 +1 overall
This message was automatically generated. |
wchevreuil
approved these changes
Nov 8, 2021
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
Apache9
approved these changes
Nov 9, 2021
There was a problem hiding this comment.
Overall LGTM. Just need to add more javadoc or comments to let later developers better understand the code.
And please fix the checkstyle issue before merging.
Thanks @BukrosSzabolcs
hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/BrokenStoreFileCleaner.java
Show resolved
Hide resolved
...main/java/org/apache/hadoop/hbase/regionserver/compactions/AbstractMultiOutputCompactor.java
Show resolved
Hide resolved
checkstyle fixes additonal comments
|
🎊 +1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
Apache9
pushed a commit
that referenced
this pull request
Nov 9, 2021
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
Apache9
pushed a commit
that referenced
this pull request
Nov 27, 2021
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
Apache9
pushed a commit
that referenced
this pull request
Dec 6, 2021
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
Apache9
pushed a commit
that referenced
this pull request
Dec 17, 2021
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
asfgit
pushed a commit
that referenced
this pull request
Dec 22, 2021
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
Apache9
pushed a commit
that referenced
this pull request
Jan 1, 2022
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
Apache9
pushed a commit
that referenced
this pull request
Jan 2, 2022
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
Apache9
pushed a commit
that referenced
this pull request
Jan 2, 2022
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
Apache9
pushed a commit
that referenced
this pull request
Jan 4, 2022
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
apurtell
pushed a commit
to apurtell/hbase
that referenced
this pull request
Mar 18, 2022
…he#3786) Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org> Conflicts: hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java hbase-server/src/main/java/org/apache/hadoop/hbase/mob/DefaultMobStoreCompactor.java
apurtell
pushed a commit
to apurtell/hbase
that referenced
this pull request
Mar 19, 2022
…he#3786) Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org> Conflicts: hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java hbase-server/src/main/java/org/apache/hadoop/hbase/mob/DefaultMobStoreCompactor.java
apurtell
pushed a commit
to apurtell/hbase
that referenced
this pull request
Mar 25, 2022
…he#3786) Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org> Conflicts: hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java hbase-server/src/main/java/org/apache/hadoop/hbase/mob/DefaultMobStoreCompactor.java
apurtell
pushed a commit
to apurtell/hbase
that referenced
this pull request
Mar 26, 2022
…he#3786) Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org> Conflicts: hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java hbase-server/src/main/java/org/apache/hadoop/hbase/mob/DefaultMobStoreCompactor.java
apurtell
added a commit
to apurtell/hbase
that referenced
this pull request
Mar 26, 2022
…others) to branch-2.5 Previous cherry picks: commit 6aaef89 HBASE-26064 Introduce a StoreFileTracker to abstract the store file tracking logic commit 43b40e9 HBASE-25988 Store the store file list by a file apache#3578) commit 6e05376 HBASE-26079 Use StoreFileTracker when splitting and merging apache#3617) commit 090b2fe HBASE-26224 HBASE-26224 Introduce a MigrationStoreFileTracker to support migratin… apache#3656) commit 0ee1689 HBASE-26246 Persist the StoreFileTracker configurations to TableDescriptor when creating table apache#3666) commit 2052e80 HBASE-26248 Should find a suitable way to let users specify the store… apache#3665) commit 5ff0f98 HBASE-26264 Add more checks to prevent misconfiguration on store file… apache#3681) commit fc4f6d1 HBASE-26280 HBASE-26280 Use store file tracker when snapshoting apache#3685) commit 06db852 HBASE-26326 CreateTableProcedure fails when FileBasedStoreFileTracker… apache#3721) commit e4e7cf8 HBASE-26386 Refactor StoreFileTracker implementations to expose the s… apache#3774) commit 08d1171 HBASE-26328 Clone snapshot doesn't load reference files into FILE SFT impl apache#3749) commit 8bec26e HBASE-26263 [Rolling Upgrading] Persist the StoreFileTracker configur… apache#3700) commit a288365 HBASE-26271: Cleanup the broken store files under data directory apache#3786) commit d00b5fa HBASE-26454 CreateTableProcedure still relies on temp dir and renames… apache#3845) commit 771e552 HBASE-26286: Add support for specifying store file tracker when restoring or cloning snapshot commit f16b7b1 HBASE-26265 Update ref guide to mention the new store file tracker im… apache#3942) commit 755b3b4 HBASE-26585 Add SFT configuration to META table descriptor when creating META apache#3998) commit 39c42c7 HBASE-26639 The implementation of TestMergesSplitsAddToTracker is pro… apache#4010) commit 6e1f5b7 HBASE-26586 Should not rely on the global config when setting SFT implementation for a table while upgrading apache#4006) commit f1dd865 HBASE-26654 ModifyTableDescriptorProcedure shoud load TableDescriptor… apache#4034) commit 8fbc9a2 HBASE-26674 Should modify filesCompacting under storeWriteLock apache#4040) commit 5aa0fd2 HBASE-26675 Data race on Compactor.writer apache#4035) commit 3021c58 HBASE-26700 The way we bypass broken track file is not enough in Stor… apache#4055) commit a8b68c9 HBASE-26690 Modify FSTableDescriptors to not rely on renaming when wr… apache#4054) commit dffeb8e HBASE-26587 Introduce a new Admin API to change SFT implementation (#… apache#4080) commit b265fe5 HBASE-26673 Implement a shell command for change SFT implementation apache#4113) commit 4cdb380 HBASE-26640 Reimplement master local region initialization to better … apache#4111) commit 77bb153 HBASE-26707: Reduce number of renames during bulkload (apache#4066) apache#4122) commit a4b192e HBASE-26611 Changing SFT implementation on disabled table is dangerous apache#4082) commit d3629bb HBASE-26837 Set SFT config when creating TableDescriptor in TestClone… apache#4226) commit 541d748 HBASE-26881 Backport HBASE-25368 to branch-2 (apache#4267) Fixups for precommit error prone, checkstyle, and javadoc warnings after applying cherry picks. Signed-off-by: Josh Elser <elserj@apache.org> Reviewed-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
apurtell
pushed a commit
that referenced
this pull request
Mar 26, 2022
Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org> Conflicts: hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HRegionServer.java hbase-server/src/main/java/org/apache/hadoop/hbase/mob/DefaultMobStoreCompactor.java
apurtell
added a commit
that referenced
this pull request
Mar 26, 2022
…others) to branch-2.5 Previous cherry picks: commit 6aaef89 HBASE-26064 Introduce a StoreFileTracker to abstract the store file tracking logic commit 43b40e9 HBASE-25988 Store the store file list by a file #3578) commit 6e05376 HBASE-26079 Use StoreFileTracker when splitting and merging #3617) commit 090b2fe HBASE-26224 HBASE-26224 Introduce a MigrationStoreFileTracker to support migratin… #3656) commit 0ee1689 HBASE-26246 Persist the StoreFileTracker configurations to TableDescriptor when creating table #3666) commit 2052e80 HBASE-26248 Should find a suitable way to let users specify the store… #3665) commit 5ff0f98 HBASE-26264 Add more checks to prevent misconfiguration on store file… #3681) commit fc4f6d1 HBASE-26280 HBASE-26280 Use store file tracker when snapshoting #3685) commit 06db852 HBASE-26326 CreateTableProcedure fails when FileBasedStoreFileTracker… #3721) commit e4e7cf8 HBASE-26386 Refactor StoreFileTracker implementations to expose the s… #3774) commit 08d1171 HBASE-26328 Clone snapshot doesn't load reference files into FILE SFT impl #3749) commit 8bec26e HBASE-26263 [Rolling Upgrading] Persist the StoreFileTracker configur… #3700) commit a288365 HBASE-26271: Cleanup the broken store files under data directory #3786) commit d00b5fa HBASE-26454 CreateTableProcedure still relies on temp dir and renames… #3845) commit 771e552 HBASE-26286: Add support for specifying store file tracker when restoring or cloning snapshot commit f16b7b1 HBASE-26265 Update ref guide to mention the new store file tracker im… #3942) commit 755b3b4 HBASE-26585 Add SFT configuration to META table descriptor when creating META #3998) commit 39c42c7 HBASE-26639 The implementation of TestMergesSplitsAddToTracker is pro… #4010) commit 6e1f5b7 HBASE-26586 Should not rely on the global config when setting SFT implementation for a table while upgrading #4006) commit f1dd865 HBASE-26654 ModifyTableDescriptorProcedure shoud load TableDescriptor… #4034) commit 8fbc9a2 HBASE-26674 Should modify filesCompacting under storeWriteLock #4040) commit 5aa0fd2 HBASE-26675 Data race on Compactor.writer #4035) commit 3021c58 HBASE-26700 The way we bypass broken track file is not enough in Stor… #4055) commit a8b68c9 HBASE-26690 Modify FSTableDescriptors to not rely on renaming when wr… #4054) commit dffeb8e HBASE-26587 Introduce a new Admin API to change SFT implementation (#… #4080) commit b265fe5 HBASE-26673 Implement a shell command for change SFT implementation #4113) commit 4cdb380 HBASE-26640 Reimplement master local region initialization to better … #4111) commit 77bb153 HBASE-26707: Reduce number of renames during bulkload (#4066) #4122) commit a4b192e HBASE-26611 Changing SFT implementation on disabled table is dangerous #4082) commit d3629bb HBASE-26837 Set SFT config when creating TableDescriptor in TestClone… #4226) commit 541d748 HBASE-26881 Backport HBASE-25368 to branch-2 (#4267) Fixups for precommit error prone, checkstyle, and javadoc warnings after applying cherry picks. Signed-off-by: Josh Elser <elserj@apache.org> Reviewed-by: Wellington Ramos Chevreuil <wchevreuil@apache.org>
vinayakphegde
pushed a commit
to vinayakphegde/hbase
that referenced
this pull request
Apr 4, 2024
…he#3786) Signed-off-by: Duo Zhang <zhangduo@apache.org> Signed-off-by: Josh Elser <elserj@apache.org> Signed-off-by: Wellington Ramos Chevreuil <wchevreuil@apache.org> Change-Id: If9825f5199385e8ccc65f4f7637461f3651a6836
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.
Add new chore to delete lefotver files in case file based storefile
handling is used
Expose the target files of currently running compactions for easier
validation