HBASE-24779 Report on the WAL edit buffer usage/limit for replication #2193
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
|
🎊 +1 overall
This message was automatically generated. |
busbey
reviewed
Aug 4, 2020
|
|
||
| @Override | ||
| public void setWALReaderEditsBufferBytes(long usage) { | ||
| //noop. Global limit, tracked globally. Do not need per-source metrics |
There was a problem hiding this comment.
nit: wouldn't it still be useful to know if particular sources were eating up the global limit? or do we not have enough information tracked already to do that?
There was a problem hiding this comment.
Once we chuck something into this usage, we have zero insight back to which source put it there. Wellington was working on this buffer tracking in HBASE-24813, but I don't think per-source tracking was "in scope"
There was a problem hiding this comment.
Once we chuck something into this usage, we have zero insight back to which source put it there.
I had the same question. I was wondering if a drill down by source would be helpful in addition to the global usage. Correct me if I'm wrong, looks like ReplicationSource class has access to the MetricsSource object.. we can just update the byte usage for that source? (and the global too at the same time). That way we can also get rid of the special logic to update one metric setWALReaderEditsBufferBytes(). Does that not work for some reason?
There was a problem hiding this comment.
looks like ReplicationSource class has access to the MetricsSource object.. we can just update the byte usage for that source? (and the global too at the same time). That way we can also get rid of the special logic to update one metric setWALReaderEditsBufferBytes()
You are correct that we could do that. I wanted to keep this change scoped on "make what we currently have reportable". I am all for doing a per-source tracking in addition to the globally-scoped tracking. I'd rather just keep these two things separate :)
There was a problem hiding this comment.
Oh, I see. Looks like you pushed the metrics update logic into the source, the patch looks clean now.
There was a problem hiding this comment.
yup! Just made another interface to isolate the new additions which cleans this up a little. I think your suggestion is still good for better, fine-grained tracking. However, since HBASE-20417, hopefully no one else runs into this ;)
hbase-server/src/test/java/org/apache/hadoop/hbase/replication/TestReplicationEndpoint.java
Show resolved
Hide resolved
hbase-server/src/test/java/org/apache/hadoop/hbase/replication/TestReplicationEndpoint.java
Outdated
Show resolved
Hide resolved
...in/java/org/apache/hadoop/hbase/replication/regionserver/MetricsReplicationSourceSource.java
Outdated
Show resolved
Hide resolved
hbase-server/src/main/java/org/apache/hadoop/hbase/replication/regionserver/Replication.java
Outdated
Show resolved
Hide resolved
|
💔 -1 overall
This message was automatically generated. |
wchevreuil
approved these changes
Aug 4, 2020
...a/org/apache/hadoop/hbase/replication/regionserver/MetricsReplicationGlobalSourceSource.java
Show resolved
Hide resolved
...c/main/java/org/apache/hadoop/hbase/replication/regionserver/ReplicationSourceWALReader.java
Outdated
Show resolved
Hide resolved
|
💔 -1 overall
This message was automatically generated. |
busbey
approved these changes
Aug 4, 2020
...a/org/apache/hadoop/hbase/replication/regionserver/MetricsReplicationGlobalSourceSource.java
Outdated
Show resolved
Hide resolved
bharathv
approved these changes
Aug 4, 2020
|
Thanks folks! Just pushed a couple more commits for cleanup on QA and wellington's suggestions. I'll merge when QA is happy. |
|
💔 -1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
|
🎊 +1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
|
Interesting. TestWALEntryStream timed out for me too. Will dig in. |
Ah, a mocking issue. Fixing. |
* Fix up metrics source interface hierarchy * Docs for test replication endpoint * Push down metrics updating to mutation of the buffer usage variable (rather than on a timer)
joshelser
force-pushed
the
24779-check-quota-insight
branch
from
e2c75b5 to
7695a70
Compare
|
I got 99 problems and all of them are due to mocking. The second UT failure was also due to a mock returning a value of 0 instead of the |
|
🎊 +1 overall
This message was automatically generated. |
|
💔 -1 overall
This message was automatically generated. |
|
TestReplicationSource failure appears to be due to HBASE-24817 (#2198) We end up filtering out an edit because it's empty, but we expect to not filter it because it's not a system table edit. I'll make the same mocking changes to prevent future pain in TestReplicationSource (like I did for TestWALEntryStream) in a follow-on, as well as fix this test. FYI @saintstack Pushing this to make some progress. |
asfgit
pushed a commit
that referenced
this pull request
Aug 7, 2020
Closes #2193 Signed-off-by: Bharath Vissapragada <bharathv@apache.org> Signed-off-by: Sean Busbey <busbey@apache.org> Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org>
|
Filed https://issues.apache.org/jira/browse/HBASE-24834 to address the TestReplicationSource failure mentioned above. |
tamasadami
pushed a commit
to tamasadami/hbase
that referenced
this pull request
Sep 9, 2020
Closes apache#2193 Signed-off-by: Bharath Vissapragada <bharathv@apache.org> Signed-off-by: Sean Busbey <busbey@apache.org> Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org> (cherry picked from commit 303db63) Change-Id: If1b3662792304747d2942dbc52338e2108b1a764
tamasadami
pushed a commit
to tamasadami/hbase
that referenced
this pull request
Sep 9, 2020
Closes apache#2193 Signed-off-by: Bharath Vissapragada <bharathv@apache.org> Signed-off-by: Sean Busbey <busbey@apache.org> Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org> (cherry picked from commit 303db63)
tamasadami
pushed a commit
to tamasadami/hbase
that referenced
this pull request
Oct 19, 2020
Closes apache#2193 Signed-off-by: Bharath Vissapragada <bharathv@apache.org> Signed-off-by: Sean Busbey <busbey@apache.org> Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org> (cherry picked from commit 303db63) Change-Id: If1b3662792304747d2942dbc52338e2108b1a764
tamasadami
pushed a commit
to tamasadami/hbase
that referenced
this pull request
Oct 19, 2020
Closes apache#2193 Signed-off-by: Bharath Vissapragada <bharathv@apache.org> Signed-off-by: Sean Busbey <busbey@apache.org> Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org> (cherry picked from commit 303db63) Change-Id: If1b3662792304747d2942dbc52338e2108b1a764
clarax
pushed a commit
to clarax/hbase
that referenced
this pull request
Nov 15, 2020
Closes apache#2193 Signed-off-by: Bharath Vissapragada <bharathv@apache.org> Signed-off-by: Sean Busbey <busbey@apache.org> Signed-off-by: Wellington Chevreuil <wchevreuil@apache.org>
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.
Testing done:
And then to observe:
replication.stats.thread.period.secondsto15in hbase-site.xml and tail regionserver logwatch -n 15 "curl -s http://mizar.local:16030/jmx | jq '.beans[] | select(.name == \"Hadoop:service=HBase,name=RegionServer,sub=Replication\") | .\"source.walReaderEditsBufferUsage\"'"Can see the buffer grow. When the puts run out, the buffer eventually drains.