HDFS-16598. Fix DataNode FsDatasetImpl lock issue without GS checks.#4366
Merged
Hexiaoqiao merged 4 commits intoapache:trunkfrom Jun 14, 2022
Merged
HDFS-16598. Fix DataNode FsDatasetImpl lock issue without GS checks.#4366Hexiaoqiao merged 4 commits intoapache:trunkfrom
Hexiaoqiao merged 4 commits intoapache:trunkfrom
Conversation
Contributor
Author
|
@Hexiaoqiao @ayushtkn This bug was introduced by HDFS-16534. Because the block GS of client may be smaller than DN when pipeline recovery failed. Please help me review it, thanks. |
This was referenced May 28, 2022
|
💔 -1 overall
This message was automatically generated. |
Contributor
|
@ZanderXu Thanks for your catches. Sorry I don't get why HDFS-16534 could lead GS inconsistency between client and datanode. Thanks again. |
Contributor
Author
|
In the original PipelineRecovery process, if the pipeline recovery failed, the block GS maybe inconsistency. So during Pipeline Recovery, GS inconsistency is expected. HDFS-16534 has a bug in handling inconsistent GS, and caused All datanodes XXX are bad. Aborting... |
Contributor
|
Thanks for your feedback.But I don't quite understand why this PR can solves the problem.It seems that you only change getReplicaInfo(x) to getReplicaInfo(x,x), the logic in two method was almost same. |
Contributor
Author
|
getReplicaInfo(ExtendedBlock b) will check gs, and getReplicaInfo(String bpid, long blkid) will not check the gs. |
Contributor
|
@ZanderXu LGTM. When we acquire volume lock we should invoke getReplicaInfo(..) to get volume uuid for replica.So the GS is not necessary check for acquire lock stage.The original method will do check GS If GS check is necessary. |
Contributor
Author
|
@MingXiangLi Thanks for you review. |
Contributor
@ZanderXu @MingXiangLi |
Contributor
Author
|
Thanks @slfan1989 for your comment. |
Contributor
@ZanderXu Thanks for the great catch here.
It is good question. |
Contributor
Author
@Hexiaoqiao do you mean to change all getReplicaInfo(ExtendedBlock b) to getReplicaInfo(String bpid, long blkid) in fine-grained lock in this PR? |
Contributor
Sure. It is not necessary to check GS when acquire BP/VOLUME lock which is totally not related with GS IMO. Thanks. |
Contributor
Author
|
Got, I will do it. |
Contributor
Author
@Hexiaoqiao @MingXiangLi @slfan1989 I have update the patch, please help me review it. Thanks About this jira, please refer to UT |
Hexiaoqiao
reviewed
Jun 11, 2022
| return info; | ||
| } | ||
|
|
||
| ReplicaInfo getReplicaInfoForLock(ExtendedBlock b) |
Contributor
There was a problem hiding this comment.
Great progress here. How about to integrate getReplicaInfoForLock ang getStorageUuid together to getStorageUuidForLock?
String getStorageUuidForLock(ExtendedBlock b)
throws ReplicaNotFoundException {
return getReplicaInfo(b.getBlockPoolId(), b.getBlockId()).getStorageUuid();
}
Contributor
Author
There was a problem hiding this comment.
Update the patch, please review again.
…ther to getStorageUuidForLock
|
💔 -1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
added 2 commits
June 12, 2022 18:37
…ataset.impl.TestWriteToReplica.testAppend
Contributor
|
@ZanderXu thanks for your contribution. |
Contributor
|
Yetus reported one failed unit test, |
|
💔 -1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
Contributor
Author
|
@Hexiaoqiao he The failed UT |
Hexiaoqiao
approved these changes
Jun 14, 2022
Contributor
Hexiaoqiao
left a comment
Hexiaoqiao
left a comment
There was a problem hiding this comment.
LGTM. +1. Will commit if no other more comments. cc @MingXiangLi
Contributor
|
LGTM.@ZanderXu thanks for your contribution. |
Hexiaoqiao
changed the title
Contributor
|
Committed to trunk. Javadoc issue reported by Yetus is not related to this PR. Which is tracking by another ticket. |
Contributor
Author
|
Thanks @Hexiaoqiao @MingXiangLi and @slfan1989 for your reviews. |
HarshitGupta11
pushed a commit
to HarshitGupta11/hadoop
that referenced
this pull request
Nov 28, 2022
…pache#4366). Contributed by ZanderXu. Reviewed-by: Mingxiang Li <liaiphag0@gmail.com> Signed-off-by: He Xiaoqiao <hexiaoqiao@apache.org>
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
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.
Detail info please refer HDFS-16598.
All datanodes [DatanodeInfoWithStorage[127.0.0.1:57448,DS-1b5f7e33-a2bf-4edc-9122-a74c995a99f5,DISK]] are bad. Aborting...