-
Bug
-
Resolution: Fixed
-
P4
-
19, repo-loom
-
b08
Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
---|---|---|---|---|---|---|
JDK-8294060 | 17.0.6-oracle | Dukebot | P4 | Resolved | Fixed | b01 |
JDK-8294218 | 17.0.6 | Goetz Lindenmaier | P4 | Resolved | Fixed | b01 |
resourcehogs/serviceability/sa/ClhsdbRegionDetailsScanOopsForG1.java is failing in the loom repo, even when not being run using the vthread wrapper.
java.lang.RuntimeException: Test ERROR java.lang.RuntimeException: 'Survivor' missing from stdout/stderr
The test runs the g1regiondetails clhsdb command, which dumps out all of the g1 regions. There's a very large number of "free" regions, 1 "Eden" region, and a couple of "Old" regions, but no "Survivor" region, which is what the test is looking for.
At first I though it may have been missed in the truncated part of the output, since the test exceeds jteg's max output, but I ran with an increased max output so there would be no truncation, and there is still no sign of a Survivor region in the output.
This failure does not happen when no JVM args are specified, but does with the following:
-Xcomp -XX:+CreateCoredumpOnCrash -ea -esa -XX:CompileThreshold=100 -XX:+UnlockExperimentalVMOptions -server -XX:+TieredCompilation
I suspect -Xcomp is what matters here, but have not verified that yet.
java.lang.RuntimeException: Test ERROR java.lang.RuntimeException: 'Survivor' missing from stdout/stderr
The test runs the g1regiondetails clhsdb command, which dumps out all of the g1 regions. There's a very large number of "free" regions, 1 "Eden" region, and a couple of "Old" regions, but no "Survivor" region, which is what the test is looking for.
At first I though it may have been missed in the truncated part of the output, since the test exceeds jteg's max output, but I ran with an increased max output so there would be no truncation, and there is still no sign of a Survivor region in the output.
This failure does not happen when no JVM args are specified, but does with the following:
-Xcomp -XX:+CreateCoredumpOnCrash -ea -esa -XX:CompileThreshold=100 -XX:+UnlockExperimentalVMOptions -server -XX:+TieredCompilation
I suspect -Xcomp is what matters here, but have not verified that yet.
- backported by
-
JDK-8294060 resourcehogs/serviceability/sa/ClhsdbRegionDetailsScanOopsForG1.java can fail if GC is triggered
- Resolved
-
JDK-8294218 resourcehogs/serviceability/sa/ClhsdbRegionDetailsScanOopsForG1.java can fail if GC is triggered
- Resolved
- links to
-
Commit openjdk/jdk17u-dev/6e07afa2
-
Commit openjdk/jdk/4532c3a1
-
Review openjdk/jdk17u-dev/712
-
Review openjdk/jdk/7257
(1 links to)