Github user srdo commented on a diff in the pull request:
https://github.com/apache/storm/pull/2113#discussion_r120444024
--- Diff: storm-server/src/main/java/org/apache/storm/DaemonConfig.java ---
@@ -889,10 +890,83 @@
public static String STORM_CGROUP_CGEXEC_CMD =
"storm.cgroup.cgexec.cmd";
/**
- * The amount of memory a worker can exceed its allocation before
cgroup will kill it
+ * Please use STORM_SUPERVISOR_MEMORY_LIMIT_TOLERANCE_MARGIN_MB
instead. The amount of memory a
+ * worker can exceed its allocation before cgroup will kill it.
+ */
+ @isPositiveNumber(includeZero = true)
+ public static String STORM_CGROUP_MEMORY_LIMIT_TOLERANCE_MARGIN_MB =
+ "storm.cgroup.memory.limit.tolerance.margin.mb";
+
+ /**
+ * Java does not always play nicely with cgroups. It is coming but not
fully implemented and not
+ * for the way storm uses cgroups. In the short term you can disable
the hard memory enforcement
+ * by cgroups and let the supervisor handle shooting workers going
over their limit in a kinder
+ * way.
+ */
+ @isBoolean
+ public static String STORM_CGROUP_MEMORY_ENFORCEMENT_ENABLE =
"storm.cgroup.memory.enforcement.enable";
+
+ // Configs for memory enforcement does by the supervisor (not cgroups
directly)
--- End diff --
nit: does -> done(?)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
