New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
KYLIN-4081 Use absolute path instead of relative for local segment cache #741
Conversation
Codecov Report
@@ Coverage Diff @@
## master #741 +/- ##
============================================
- Coverage 25.71% 25.71% -0.01%
Complexity 6021 6021
============================================
Files 1387 1387
Lines 82758 82760 +2
Branches 11586 11586
============================================
- Hits 21282 21279 -3
- Misses 59432 59436 +4
- Partials 2044 2045 +1
Continue to review full report at Codecov.
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
If user specify an absolute path, such as /path1/index, should respect that, otherwise if it is an relative path, use KylinHome as parent
Yes, you are right. |
1. Use absolute path instead of relative path for local segment cache may cause unpredicted result. After restart receiver, if working directory change, Receiver cannot find segment data/checkpoint which created previously, and recreate new one, which is a serious problem. 2. Add several way of metrics for streaming receiver, including console/csv/jmx.
Test Local Segment Cache Path
All manual test passed. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Test Metricskylin.stream.metrics.option=consoleYou can get metrics in logs/streaming_receiver.out.
kylin.stream.metrics.option=csvYou can get all file under $KYLIN_HOME/stream_metrics_csv, like this [root@cdh-client apache-kylin-3.0.0-SNAPSHOT-bin-4081-receiver]# tree stream_metrics_csv/
stream_metrics_csv/
├── gc.PS-MarkSweep.count.csv
├── gc.PS-MarkSweep.time.csv
├── gc.PS-Scavenge.count.csv
├── gc.PS-Scavenge.time.csv
├── memory.heap.committed.csv
├── memory.heap.init.csv
├── memory.heap.max.csv
├── memory.heap.usage.csv
├── memory.heap.used.csv
├── memory.non-heap.committed.csv
├── memory.non-heap.init.csv
├── memory.non-heap.max.csv
├── memory.non-heap.usage.csv
├── memory.non-heap.used.csv
├── memory.pools.Code-Cache.committed.csv
├── memory.pools.Code-Cache.init.csv
├── memory.pools.Code-Cache.max.csv
├── memory.pools.Code-Cache.usage.csv
├── memory.pools.Code-Cache.used.csv
├── memory.pools.Compressed-Class-Space.committed.csv
├── memory.pools.Compressed-Class-Space.init.csv
├── memory.pools.Compressed-Class-Space.max.csv
├── memory.pools.Compressed-Class-Space.usage.csv
├── memory.pools.Compressed-Class-Space.used.csv
├── memory.pools.Metaspace.committed.csv
├── memory.pools.Metaspace.init.csv
├── memory.pools.Metaspace.max.csv
├── memory.pools.Metaspace.usage.csv
├── memory.pools.Metaspace.used.csv
├── memory.pools.PS-Eden-Space.committed.csv kylin.stream.metrics.option=jmx |
Use absolute path instead of relative path for local segment cache may cause unpredicted result. After restart receiver, if working directory change, Receiver cannot find segment data/checkpoint which created previously, and recreate new one, which is a serious problem.