From 6358388ad9f7676aa3b1ab149c9fbde4716929e5 Mon Sep 17 00:00:00 2001 From: Peter Amstutz Date: Tue, 19 Apr 2022 15:40:56 -0400 Subject: [PATCH] 17301: Report warning about OOM killer when exit code 137 Arvados-DCO-1.1-Signed-off-by: Peter Amstutz --- sdk/cwl/arvados_cwl/arvcontainer.py | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sdk/cwl/arvados_cwl/arvcontainer.py b/sdk/cwl/arvados_cwl/arvcontainer.py index e2c2f2e67b..c85443a23a 100644 --- a/sdk/cwl/arvados_cwl/arvcontainer.py +++ b/sdk/cwl/arvados_cwl/arvcontainer.py @@ -392,6 +392,10 @@ class ArvadosContainer(JobBase): processStatus = "success" else: processStatus = "permanentFail" + + if rcode == 137: + logger.warning("%s job was killed on the compute instance. The most common reason is that it attempted to allocate too much RAM and was targeted by the Out Of Memory (OOM) killer. Try resubmitting with a higher 'ramMin'.", + self.arvrunner.label(self)) else: processStatus = "permanentFail" -- 2.30.2