*/: add pd write frequency control from cdc server parameter #937
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
amyangfei
force-pushed
the
pd-update-frequency
branch
from
d9edc89
to
8af91c5
Compare
amyangfei
added
the
release-blocker
|
please resolve the conflicts |
amyangfei
force-pushed
the
pd-update-frequency
branch
2 times, most recently
from
491b342
to
6a82f78
Compare
overvenus
approved these changes
Sep 9, 2020
Comment on lines
+118
to
+125
| metrics := map[string]prometheus.Counter{ | ||
| etcd.EtcdPut: etcdRequestCounter.WithLabelValues(etcd.EtcdPut, captureAddr), | ||
| etcd.EtcdGet: etcdRequestCounter.WithLabelValues(etcd.EtcdGet, captureAddr), | ||
| etcd.EtcdDel: etcdRequestCounter.WithLabelValues(etcd.EtcdDel, captureAddr), | ||
| etcd.EtcdTxn: etcdRequestCounter.WithLabelValues(etcd.EtcdTxn, captureAddr), | ||
| etcd.EtcdGrant: etcdRequestCounter.WithLabelValues(etcd.EtcdGrant, captureAddr), | ||
| etcd.EtcdRevoke: etcdRequestCounter.WithLabelValues(etcd.EtcdRevoke, captureAddr), | ||
| } |
There was a problem hiding this comment.
IMO, it's better to put the map in package pkg/etcd, and pass capture address to the Wrap.
There was a problem hiding this comment.
the metrics has cdc specific label capture, so I think it is better to put the metrics map in the upper caller.
amyangfei
force-pushed
the
pd-update-frequency
branch
from
1895ce3
to
bd05ba7
Compare
zier-one
approved these changes
Sep 9, 2020
ti-srebot
approved these changes
Sep 9, 2020
|
please resolve the conflicts |
amyangfei
force-pushed
the
pd-update-frequency
branch
from
bd05ba7
to
69a43d1
Compare
|
/run-integration-tests tidb=release-4.0 tikv=release-4.0 pd=release-4.0 |
1 similar comment
|
/run-integration-tests tidb=release-4.0 tikv=release-4.0 pd=release-4.0 |
- p.position.CHeckPointTs update to latest checkpoint, but interval doesn't reach flushCheckpointInterval, doesn't flush - checkpoint doesn't push, such as waiting for DDL execute, the checkpoint will never forward then.
amyangfei
force-pushed
the
pd-update-frequency
branch
from
025f98c
to
3bcb37f
Compare
|
/run-integration-tests tidb=release-4.0 tikv=release-4.0 pd=release-4.0 |
Codecov Report
@@ Coverage Diff @@
## master #937 +/- ##
===========================================
Coverage 33.4140% 33.4140%
===========================================
Files 100 100
Lines 11974 11974
===========================================
Hits 4001 4001
Misses 7581 7581
Partials 392 392 |
|
/run-integration-tests tidb=release-4.0 tikv=release-4.0 pd=release-4.0 |
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.
What problem does this PR solve?
Currently, we use a watch mechanism to update processor/changefeed checkpoint, the update frequency could be 20+/s, which may be too frequent in most scenarios.
What is changed and how it works?
Add parameters in cdc server to control owner/processor flush checkpoint interval
Before, with 200ms interval and 500ms interval
After resolved ts check, processor flush interval=100ms, owner flush interval=200ms. But this has a bug, if both checkpoint ts and resolved ts are not forwarded and new table is dispatched to this processor, replication will be blocked!. We should fix this.
Current status
Check List
Tests
Related changes
Release note