Repository navigation
[NO-SNOW] Make CI Happy - #1167
Conversation
…generate Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
Trial: port JDBC 3.26 SNOW-2226852 so InsertRowsBenchmarkTest does not sit 30s on idle non-daemon uploaders and get force-killed mid JaCoCo dump.
FlushRows was a manual readback compare, not a CI canary. JUnit @parameterized never reached the JMH fork, so InsertRows always benched FDN twice; @PARAM makes the child JVM actually run FDN and Iceberg.
channel.close() was not throwing in CI; the JMH hang was TelemetryThreadPool.
…h works Iceberg JMH @PARAM hits ChannelCache.setNeedFlush; without addChannel the table is missing and launchBenchmark fails on every cloud.
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #1167 +/- ##
==========================================
+ Coverage 55.82% 60.89% +5.06%
==========================================
Files 214 214
Lines 14616 14674 +58
Branches 1606 1614 +8
==========================================
+ Hits 8160 8936 +776
+ Misses 5859 5042 -817
- Partials 597 696 +99 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
| * allowing the pool to shrink back to zero during periods of inactivity. | ||
| */ | ||
| private TelemetryThreadPool() { | ||
| // Daemon so client.close() → telemetry.flush does not pin the process for keepAliveTime. |
There was a problem hiding this comment.
can you link me the file in the jdbc -- want to make sure this is exactly what jdbc is dong
There was a problem hiding this comment.
the telemetry pr that introduced the daemon fix: snowflakedb/snowflake-jdbc#2297
what is currently in telemetry's repo: https://github.com/snowflakedb/snowflake-jdbc/blob/56e500b5e791ebbc9c17733a0b951abfa86a89b8/src/main/java/net/snowflake/client/internal/jdbc/telemetryOOB/TelemetryThreadPool.java#L4
both of them are exactly the same as this pr's backport
Summary