HDDS-3205. DeleteBlock via Ratis in SCM HA #1780

runzhiwang · 2021-01-12T06:44:30Z

What changes were proposed in this pull request?

DeleteBlock needs to add/update/delete DELETED_BLOCKS_TABLE in SCMMetadataDB. So SCM does these operations via ratis.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-3205

How was this patch tested?

Existed UT.

runzhiwang · 2021-01-12T06:45:23Z

@GlenGeng @amaliujia Could you help review this ? Thanks a lot.

amaliujia · 2021-01-12T06:48:38Z

Will take a look! Thanks for this PR.

...s/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManager.java

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java

hadoop-hdds/interface-server/src/main/proto/ScmServerDatanodeHeartbeatProtocol.proto

GlenGeng-awx

LGTM. left some inline comments. Thanks for the work!

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/BlockManagerImpl.java

...rver-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManagerImpl.java

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java

hadoop-hdds/interface-server/src/main/proto/ScmServerDatanodeHeartbeatProtocol.proto

...s/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManager.java

...rver-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManagerImpl.java

hadoop-hdds/interface-server/src/main/proto/SCMRatisProtocol.proto

runzhiwang · 2021-01-22T02:43:21Z

@GlenGeng @bshashikant @amaliujia Thanks for review. I have updated the patch.

GlenGeng-awx · 2021-01-24T16:00:52Z

@runzhiwang Is the CI failure related to your changes ?

They seems to be affected by your changes.

TestStorageContainerManager.testBlockDeletingThrottling:382
TestStorageContainerManager.testBlockDeletionTransactions:290
TestDeleteWithSlowFollower.testDeleteKeyWithSlowFollower:282 ? Timeout Timed o…

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/ha/SCMStateMachine.java

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLog.java

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/ha/SCMContext.java

hadoop-hdds/interface-server/src/main/proto/ScmServerDatanodeHeartbeatProtocol.proto

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/BlockManagerImpl.java

...s/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManager.java

...rver-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManagerImpl.java

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/ha/SCMStateMachine.java

GlenGeng-awx

LGTM, some typo related comments.

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/BlockManagerImpl.java

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java

...rver-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManagerImpl.java

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java

...rver-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManagerImpl.java

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/ha/SCMDBTransactionBuffer.java

...rver-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManagerImpl.java

GlenGeng-awx · 2021-01-27T11:00:03Z

LGTM.

@bshashikant @amaliujia @nandakumar131 Please take one more look at this PR. Thanks!

This PR has implemented all the functionality according to the design: https://docs.google.com/document/d/166Aea2EowSGWtAFWNlDv0gu4rA06dQ2rJAsBd-l210Q/edit?usp=sharing.

After it is merged. all the features of Phase 2.0 of SCM HA will be finished.

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java

amaliujia · 2021-01-28T00:05:00Z

LGTM

GlenGeng-awx · 2021-01-28T02:47:11Z

@runzhiwang Please rebase the HDDS-2823, since SCMServiceManager is merged.

runzhiwang · 2021-01-28T02:58:09Z

@GlenGeng I have rebased it.

bshashikant · 2021-01-28T16:22:29Z

Thanks @runzhiwang for the contribution.

linyiqun · 2021-02-12T15:33:43Z

Sorry for the late comments.

I have one concern about current SCM HA handling mechanism. As the Blocks/Container/Pipeline update ops are cached in BatchOperation, and only be flushed during SCM snapshot take. But if user use the stand alone SCM mode to run based on current HA code. Is that will lead OOM error since stand alone SCM won't take any snapshot by itself while the request update will still go into SCM Transaction Buffer in our current implementation.

@runzhiwang @GlenGeng , is this a real problem of my concern?

GlenGeng-awx · 2021-02-22T03:08:21Z

Hey @linyiqun, thanks for the review !

We planed to use single server raft cluster for the stand alone mode, which means the taking snapshot functionality works too.

linyiqun · 2021-02-22T05:33:35Z

We planed to use single server raft cluster for the stand alone mode, which means the taking snapshot functionality works too.

Okay, after that, that's not a problem.

runzhiwang changed the title ~~HDDS-3205. Handle BlockDeletingService in SCM HA~~ HDDS-3205. DeleteBlock via Ratis in SCM HA Jan 12, 2021

runzhiwang closed this Jan 13, 2021

runzhiwang reopened this Jan 13, 2021

amaliujia reviewed Jan 13, 2021

View reviewed changes

...s/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogStateManager.java Outdated Show resolved Hide resolved

amaliujia reviewed Jan 13, 2021

View reviewed changes

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java Outdated Show resolved Hide resolved

bshashikant reviewed Jan 13, 2021

View reviewed changes

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java Outdated Show resolved Hide resolved

bshashikant reviewed Jan 13, 2021

View reviewed changes

hadoop-hdds/interface-server/src/main/proto/ScmServerDatanodeHeartbeatProtocol.proto Outdated Show resolved Hide resolved

GlenGeng-awx reviewed Jan 13, 2021

View reviewed changes

runzhiwang force-pushed the HDDS-2823-delete-block branch from 31daaec to 565326f Compare January 15, 2021 08:26

bshashikant reviewed Jan 15, 2021

View reviewed changes

hadoop-hdds/interface-server/src/main/proto/SCMRatisProtocol.proto Outdated Show resolved Hide resolved

mukul1987 added the scm-ha label Jan 19, 2021

runzhiwang force-pushed the HDDS-2823-delete-block branch from 565326f to 9de897b Compare January 22, 2021 01:34

GlenGeng-awx reviewed Jan 24, 2021

View reviewed changes

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/ha/SCMStateMachine.java Outdated Show resolved Hide resolved

GlenGeng-awx reviewed Jan 24, 2021

View reviewed changes

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLog.java Outdated Show resolved Hide resolved

GlenGeng-awx reviewed Jan 24, 2021

View reviewed changes

hadoop-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/ha/SCMContext.java Outdated Show resolved Hide resolved

GlenGeng-awx reviewed Jan 25, 2021

View reviewed changes

runzhiwang closed this Jan 26, 2021

runzhiwang reopened this Jan 26, 2021

GlenGeng-awx reviewed Jan 27, 2021

View reviewed changes

runzhiwang force-pushed the HDDS-2823-delete-block branch 2 times, most recently from d62e6cc to 744f6e3 Compare January 27, 2021 08:59

amaliujia reviewed Jan 27, 2021

View reviewed changes

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java Outdated Show resolved Hide resolved

...op-hdds/server-scm/src/main/java/org/apache/hadoop/hdds/scm/block/DeletedBlockLogImplV2.java Outdated Show resolved Hide resolved

HDDS-3205. Handle BlockDeletingService in SCM HA

928a9ce

runzhiwang force-pushed the HDDS-2823-delete-block branch from 744f6e3 to 928a9ce Compare January 28, 2021 02:56

bshashikant approved these changes Jan 28, 2021

View reviewed changes

bshashikant merged commit b328608 into apache:HDDS-2823 Jan 28, 2021

HDDS-3205. DeleteBlock via Ratis in SCM HA #1780

HDDS-3205. DeleteBlock via Ratis in SCM HA #1780

Uh oh!

Conversation

runzhiwang commented Jan 12, 2021

What changes were proposed in this pull request?

What is the link to the Apache JIRA

How was this patch tested?

Uh oh!

runzhiwang commented Jan 12, 2021

Uh oh!

amaliujia commented Jan 12, 2021

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GlenGeng-awx left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

runzhiwang commented Jan 22, 2021

Uh oh!

GlenGeng-awx commented Jan 24, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GlenGeng-awx left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

GlenGeng-awx commented Jan 27, 2021

Uh oh!

Uh oh!

Uh oh!

amaliujia commented Jan 28, 2021

Uh oh!

GlenGeng-awx commented Jan 28, 2021

Uh oh!

runzhiwang commented Jan 28, 2021

Uh oh!

bshashikant commented Jan 28, 2021

Uh oh!

linyiqun commented Feb 12, 2021

Uh oh!

GlenGeng-awx commented Feb 22, 2021

Uh oh!

linyiqun commented Feb 22, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

GlenGeng-awx commented Jan 24, 2021 •

edited

Loading