Closed thinker0 closed 2 years ago
https://app.travis-ci.com/github/apache/incubator-heron/builds/248849844 How do I solve this ?
grep: write error: No space left on device
https://app.travis-ci.com/github/apache/incubator-heron/builds/248849844 How do I solve this ?
I've restarted the build.
I have restarted the build to see if the build errors are transient. I do not see a reason why updates/changes to logging should cause test failures.
Something is causing the creation of large files in this PR. Since this PR is making changes to the logger I am wondering if the log files are getting too large. The actual production code seems to be fine but there may need to be changes in the log output or log file size.
OSError: [Errno 28] No space left on device
I do not believe this is a build artifact size related issue because I think that would have triggered earlier after build completion.
I have tried to rerun the pipeline three times to see if there was a load issue causing this with Travis CI but they fail in roughly the same place.
There is an issue on SO that suggests this occurs with Python3 and could be due to either a lack of disk space in /tmp
or due too many files. Are the log file structures different from before? Are more or larger logs being written?
Nice ^____^;
Just one final concern/question: Can the log file size/count affect production systems, and if so, how might we mitigate the issue now?
@surahman
jul-to-slf4j : Default FINE
slf4j-to-log4j: Default INFO ?
logback: Default DEBUG
As expected, it seems that it is normal to have less than before. The definition in logback.xml is INFO Level, so some contents should be reduced.
logback.xml: INFO
heron % ./scripts/travis/ci.sh 2>&1 | tee ci.log
===========================================================
heron build integration_test 0:00:31
heron install 0:00:22
heron tests install 0:00:06
heron integration_test local 0:04:32
heron integration_test http-server initialization 0:00:00
heron integration_test scala 0:01:12
heron integration_test java 0:12:08
heron integration_test python 0:03:22
heron integration_topology_test java 0:02:49
===> Finished ./scripts/travis/test.sh at 2022-04-06 11:27:40 (0:25:02)
===> Task duration summary for ./scripts/travis/ci.sh
===========================================================
./scripts/travis/build.sh 0:26:35
./scripts/travis/test.sh 0:25:02
ls -la -rt
-rw-r--r-- 1 thinker0 staff 2.5M Apr 6 10:56 heron_build.txt
-rw-r--r-- 1 thinker0 staff 4.3K Apr 6 10:58 heron_test_non_flaky.txt
-rw-r--r-- 1 thinker0 staff 1.7K Apr 6 10:59 heron_test_flaky.txt
-rw-r--r-- 1 thinker0 staff 2.0K Apr 6 11:01 heron_build_tarpkgs.txt
-rw-r--r-- 1 thinker0 staff 1.1K Apr 6 11:02 heron_build_binpkgs.txt
-rw-r--r-- 1 thinker0 staff 3.2K Apr 6 11:03 heron_build_integration_test.txt
-rw-r--r-- 1 thinker0 staff 1.3K Apr 6 11:03 heron_install.txt
-rw-r--r-- 1 thinker0 staff 1.4K Apr 6 11:03 heron_tests_install.txt
drwxrwx--- 67 thinker0 staff 2.1K Apr 6 11:07 .
drwxr-xr-x 137 thinker0 staff 4.3K Apr 6 11:24 results
-rw-r--r-- 1 thinker0 staff 429K Apr 6 11:27 ci.log
When I test, the log file is not large.
@nicknezis requested more reviews so I have not merged the PR. I think everything is in order.
Edit: The reason I was looking at the log file size or count was based on the context of the PR and the Python3 error.
Adding for context, this was a further refinement on #3807 . Thanks @thinker0 !
JavaUtilLogging to SLF4JBridgeHandler