Scott Lystig Fritchie
Scott Lystig Fritchie
I don't have a lot of experience with the `setrlimit(3)` categories that are relevant/overlap with the intent of the limit in this PR, e.g., https://linux.die.net/man/2/setrlimit with RLIMIT_DATA, RLIMIT_RSS, and RLIMIT_STACK....
In the weekly Pony meeting today, @SeanTAllen wondered if existing setrlimit and perhaps other container'ish resource limit mechanisms were sufficient to limit memory growth in the same way that the...
I believe that I've seen this problem happen again, in the context of a `tests/system_events.py::test_autoscale_python2_MultiPartitionDetector_alo_1_Shrink1_Wait2_Shrink4` test instead of the simpler Pony `Grow1` test. The test hangs after the 2nd shrink....
I believe this is now documented, and so this ticket can remain closed.
All instances of an empty pipeline or a pipeline with a single stateless or state computation exhibit the same behavior: only a single sink TCP connection sends data to the...
This is a bad sequence: the `LogRotationResumeBarrierToken` doesn't appear in this bad case, but `ForwardedInjectBarrierFullyAckedMsg` does: ``` 1562619174.718816,INIT,Checkpoint_Initiator: Event Log CheckpointId 3 complete for worker initializer 1562619174.848105,INIT,YO: ExternalMsgDecoder._RotateLog 1562619174.848110,INIT,Received ExternalRotateLogFilesMsg...
This patch "eliminates" the test failures, until the bug can be found & fixed. ``` diff --git a/testing/tools/integration/cluster.py b/testing/tools/integration/cluster.py index 174043c72..479db7ae1 100644 --- a/testing/tools/integration/cluster.py +++ b/testing/tools/integration/cluster.py @@ -631,6 +631,7 @@...
No, this ticket should remain open. PR #2960 didn't fix the underlying problem.
No, sorry, I haven't tested it with Python3 yet.
In a recent project, I've used the following, when we knew that we wanted at least 40 source connections for a high-throughput pipeline. ``` .new_pipeline[Array[U8] val, None]( "Busy pipeline", TCPSourceConfig[Array[U8]...